The Hindu Marriage Act 1955 did two things at once that no earlier Indian statute had attempted for the majority community of the country. It declared that a Hindu marriage is a legal status created on stated conditions, monogamous, registered in a scheme of courts and decrees, and it declared that the status could be ended. For the communities it governs, marriage had been understood across centuries of shastric law as a sacrament and an indissoluble union, a bond that death itself did not sever in the texts that orthodox opinion treated as authoritative. The Act replaced that understanding with a code: five conditions for a valid marriage, a closed list of grounds on which a court may dissolve it, a set of lesser remedies for marriages in distress, and a machinery of maintenance and custody to govern the consequences. Everything in the statute flows from the tension between those two conceptions, and the statute’s largest unresolved problem flows from it too. The Act permits divorce only where one spouse can prove a listed ground, most of which require fault or a specified disability. A marriage in which neither party is at fault in any provable way, but which has ended in every sense except the legal one, has no exit in the text. That gap, between the breakdown people experience and the grounds Parliament enacted, is the subject of the longest-running unfinished argument in Indian family law, and it is where this article ends.

The article follows the statute’s own architecture. It begins with the position before 1955 and the failure that made codification unavoidable, then identifies the persons the Act binds and the concepts it defines, works through the conditions of a valid marriage and the ceremonies that complete one, and then takes the remedies in ascending order of severity: restitution of conjugal rights, judicial separation, nullity, divorce on the grounds Parliament listed, divorce by mutual consent, and the one route to dissolution that exists outside the statute altogether, in the plenary power of the Supreme Court. Maintenance, custody, forum and procedure are treated with the decrees they attach to. The assessment at the end is candid about what the Act does well and what it has never done.

What existed before the Hindu Marriage Act 1955, and why it failed

The law the Act displaced was not a single system. Classical Hindu law, drawn from the dharmashastra texts and their commentaries and administered through the two great schools of Mitakshara and Dayabhaga, treated marriage as a samskara, a sacrament that transformed the parties and could not be undone by either of them or by any court. Under that conception there was no divorce. A marriage validly solemnised endured for life and, in orthodox doctrine, beyond it, which is why the remarriage of widows stood outside the religious law until the Hindu Widows’ Remarriage Act of 1856 created a statutory exception. Polygamy was permitted to a Hindu husband by the classical law, restrained by custom in some communities and practised in others, and a wife had no corresponding liberty and no general right to leave a marriage that had failed, however brutal or empty it had become.

Against that background, the exceptions matter as much as the rule, because they showed the legislature where the pressure lay. Custom recognised divorce in a number of communities, particularly among groups the orthodox texts ranked lower in the ritual order, and the courts enforced proved custom even against the shastric rule; the Act itself would later save customary divorce expressly. Provincial legislatures, acting in the years before independence when marriage was a subject on which they could legislate, began to supply by statute what the religious law withheld. The Bombay Hindu Divorce Act of 1947, the Madras Hindu (Bigamy Prevention and Divorce) Act of 1949 and the Saurashtra Hindu Divorce Act of 1952 each introduced dissolution and, in the Madras statute, the prevention of bigamous marriages, for the Hindus of their provinces. The result by the early 1950s was a patchwork: a Hindu in Bombay could obtain a divorce that a Hindu in Bengal could not, bigamy was a statutory wrong in one province and a religious entitlement in the next, and the validity of a marriage or the legitimacy of a child could turn on geography.

Codification was the reform programme of the first Parliament, and marriage was its most contested field. The comprehensive Hindu Code Bill, prepared under B. R. Ambedkar as Law Minister and built on the work of the Hindu Law Committee chaired by B. N. Rau, proposed to restate the whole of Hindu personal law, marriage, succession, adoption, minority and maintenance, in a single modern code. The Bill met organised resistance from orthodox opinion, from the President’s office in correspondence that has since become part of the record, and from within the governing party itself, and it stalled. Ambedkar resigned from the cabinet in 1951, and his resignation statement named the shelving of the Code among his reasons. The government of Jawaharlal Nehru then carried the programme in pieces. Four separate Acts emerged: the Hindu Marriage Act of 1955, and in 1956 the Hindu Succession Act, the Hindu Minority and Guardianship Act and the Hindu Adoptions and Maintenance Act. The contested passage, the debates over monogamy and divorce, and the distance between the Code as drafted and the Acts as passed are the subject of the account of the Hindu Code Bills and their passage, which owns the enactment story; this article takes the enacted text as its subject.

What the piecemeal route meant for the marriage statute is worth one observation, because it explains the text’s character. The Act of 1955 is a compromise instrument. It imposed monogamy as a condition of validity, which was the reformers’ central demand, and it introduced divorce, which was the demand orthodoxy resisted most; but it confined divorce to a list of fault and disability grounds drawn largely from English matrimonial law of the period, it retained the remedy of restitution of conjugal rights that English law itself would abolish within a generation, and it preserved the sacramental theory in its law of ceremonies, under which a marriage is made by the customary rites of the parties rather than by registration or civil form. The Act is therefore best read as two settlements layered together: a settlement about entry into marriage, where the reform is nearly complete, and a settlement about exit, where the reform was deliberately partial and has been contested ever since.

The drafters of 1955 did not invent the enumerated fault list; they inherited it. The Indian Divorce Act of 1869 had governed Christian marriages for nearly a century on the model of the English matrimonial causes legislation, dissolving marriages on stated grounds proved in a district court, and the Dissolution of Muslim Marriages Act of 1939 had given Muslim wives a statutory list of grounds on which to seek dissolution from a court, the first general statutory divorce jurisdiction for a religious community in India. The Hindu Marriage Act adopted the form both statutes used: a code, a court, a list, a decree. What it could not adopt was their premise. The Christian and Muslim statutes reformed communities whose marriage law already admitted, in differing measures, that a marriage could end; the Hindu statute legislated divorce into a tradition whose authoritative texts admitted no ending at all, and the resistance that broke the comprehensive Code and delayed the marriage Act was concentrated precisely on that novelty and on monogamy. The compromise visible in the finished text, divorce confined to grave fault and disability, restitution retained, ceremonies left to custom, is the record of that resistance negotiated clause by clause, and the genealogy explains the otherwise puzzling fact that the Act of 1955 was at once the most radical marriage statute in the country’s history and, in its law of exit, more conservative than the statute Parliament had given Muslim wives sixteen years earlier.

Three features of the pre-1955 position survived into the Act and recur throughout this article. The first is the authority of custom. The Act defines custom and usage as a rule continuously and uniformly observed for a long time, which has obtained the force of law among Hindus in a local area, tribe, community, group or family, provided it is certain, not unreasonable and not opposed to public policy, and it lets custom displace two of the five conditions of validity and supply ceremonies and even a mode of divorce. The second is the sacramental form: a Hindu marriage is solemnised, not contracted before an officer, and its validity turns on rites rather than paperwork, which is why registration occupies the ambiguous position examined below. The third is inequality of exit. The classical law gave the husband freedoms it denied the wife; the Act formally equalised the grounds between spouses, then added four grounds available to a wife alone, in recognition of the specific wrongs the older law had licensed. Each of these survivals is a decision the drafters made with their eyes open, and each still does work in the courts.

Who the Hindu Marriage Act 1955 binds, and the concepts it defines

The Act’s long title describes it as an Act to amend and codify the law relating to marriage among Hindus, and it is Act 25 of 1955. It received the assent of the President on 18 May 1955 and came into force on the same day. Its application clause, Section 2, is drafted more widely than its title suggests. The Act applies to any person who is a Hindu by religion in any of its forms or developments, expressly including a Virashaiva, a Lingayat and a follower of the Brahmo, Prarthana or Arya Samaj; to any person who is a Buddhist, a Jain or a Sikh by religion; and to any other person domiciled in the territories to which the Act extends who is not a Muslim, a Christian, a Parsi or a Jew by religion, unless it is proved that such a person would not have been governed by Hindu law or by custom as part of that law in the matters the Act covers. Converts to any of these religions are within the clause, and a child both of whose parents, or one of whose parents, is a Hindu, Buddhist, Jain or Sikh is treated as belonging to that religion for the purposes of the Act where the child is brought up as a member of the community concerned. The Act applies also to Hindus domiciled in India who are outside its territory.

Two exclusions define the boundary. The Act does not apply to members of a Scheduled Tribe within the meaning of the Constitution unless the Central Government directs otherwise by notification, a clause that preserves tribal customary marriage law across large parts of the country and that litigants overlook at their cost. And the Act governs marriage between two Hindus: where one party is not within Section 2, the statute does not supply the law of the marriage, which is the point at which the civil alternative considered later in this article enters. Section 4 gives the Act overriding effect: any text, rule or interpretation of Hindu law, and any custom or usage, ceases to have effect on any matter for which the Act makes provision, and any other law in force ceases to apply to Hindus so far as it is inconsistent with the Act. The codification is therefore genuine, subject to the savings the Act itself records, including the saving in Section 29 for rights of divorce recognised by custom and for marriages dissolved under special enactments before the Act commenced.

The definitions in Section 3 carry more weight than definitions usually do, because two of the five conditions of validity are expressed through them. The degrees of prohibited relationship are defined by a list of relationships: two persons are within the degrees if one is a lineal ascendant of the other, if one was the wife or husband of a lineal ascendant or descendant of the other, if one was the wife of the brother or of the father’s or mother’s brother or of the grandfather’s or grandmother’s brother of the other, or if the two are brother and sister, uncle and niece, aunt and nephew, or children of brother and sister or of two brothers or of two sisters. Sapinda relationship is defined by ascent: two persons are sapindas of each other if one is a lineal ascendant of the other within the limits of sapinda relationship, or if they have a common lineal ascendant who is within those limits for each of them, the relationship extending as far as the third generation inclusive in the line of ascent through the mother and the fifth generation inclusive in the line of ascent through the father, the line being traced upwards in each case from the person concerned as the first generation. These definitions look forbidding on the page and are simple in operation: they mark out a circle of kin, fixed by counting generations, inside which marriage is forbidden unless custom permits it, and the circle differs across India because the customs that displace it differ, most famously in the southern cross-cousin and uncle and niece marriages that custom sanctions and the Act therefore allows.

The forum the Act creates is the district court, which in practice means the family court where the Family Courts Act of 1984 has established one. Petitions under the Act are presented to the district court within whose local limits the marriage was solemnised, or the respondent resides at the date of presentation, or the parties to the marriage last resided together, or, in a defined case where the respondent is outside the territories to which the Act extends or has not been heard of as alive for seven years, where the petitioner resides. A wife who petitions may additionally file where she is residing on the date of presentation, a clause inserted by the Marriage Laws (Amendment) Act of 2003, Act 50 of 2003, which corrected a structural disadvantage under which a wife who had left the matrimonial home had to sue where her husband chose to live. Proceedings are held in camera when either party so desires or the court so thinks fit, and it is unlawful to print or publish the substance of such proceedings except a judgment of a High Court or the Supreme Court with the previous permission of the court, a provision that gives matrimonial litigation a privacy the general run of civil litigation lacks.

The five conditions of a valid Hindu marriage

Section 5 states the conditions on which a marriage may be solemnised between two Hindus, and it is the section on which the validity of every marriage under the Act ultimately depends. A marriage may be solemnised if neither party has a spouse living at the time of the marriage; if neither party is incapable of giving a valid consent to it in consequence of unsoundness of mind, or, though capable of giving a valid consent, has been suffering from mental disorder of such a kind or to such an extent as to be unfit for marriage and the procreation of children, or has been subject to recurrent attacks of insanity; if the bridegroom has completed the age of twenty-one years and the bride the age of eighteen years at the time of the marriage; if the parties are not within the degrees of prohibited relationship, unless the custom or usage governing each of them permits of a marriage between the two; and if the parties are not sapindas of each other, unless the custom or usage governing each of them permits of a marriage between the two. Five conditions, but they are not of one kind, and the differences among them drive the nullity law examined in the next sections.

The first condition is monogamy, and it is the Act’s sharpest break with the law it replaced. The classical law permitted a Hindu husband more than one wife; Section 5(i) makes a living spouse an absolute bar, Section 11 declares a marriage solemnised in contravention of the condition null and void, and Section 17 applies the punishment provisions of the Indian Penal Code on bigamy, Sections 494 and 495, to a marriage that is void by reason of a spouse living. The condition cannot be displaced by custom, and it is not softened by the consent of the first wife, by the parties’ community practice, or by the conversion of either party to a religion whose personal law permits polygamy, a proposition the courts developed in the conversion litigation treated later in this article. Monogamy under the Act is not a default rule; it is the price of the statute’s recognition.

The second condition governs mental capacity, and its drafting history is a small lesson in how the Act has been maintained. As enacted in 1955, the condition spoke of neither party being an idiot or a lunatic at the time of the marriage, language borrowed from an older psychiatry and an older insensitivity. The Marriage Laws (Amendment) Act of 1976, Act 68 of 1976, substituted the present three-limbed clause: incapacity to give valid consent in consequence of unsoundness of mind; mental disorder of such a kind or to such an extent as to render the party unfit for marriage and the procreation of children, though the party is capable of giving a valid consent; and recurrent attacks of insanity. The substitution matters beyond its vocabulary. The 1955 text asked whether a party fell within a status; the 1976 text asks functional questions about consent and fitness, and the same functional language was carried into the divorce ground on mental condition in Section 13(1)(iii), so that condition and ground are read together. A further maintenance followed in 1999, when the Marriage Laws (Amendment) Act of that year, Act 39 of 1999, deleted the reference to epilepsy that had stood in the third limb beside insanity, removing a diagnosis from the statute book on the ground that its presence stigmatised a treatable neurological condition. Where a marriage is solemnised in breach of the capacity condition, it is not void but voidable at the instance of the other party under Section 12, a distinction whose consequences are drawn out below.

The third condition is age. As enacted, Section 5(iii) required the bridegroom to have completed eighteen years and the bride fifteen, figures that recorded the reformers’ defeat as much as their victory, since the Age of Consent and child marriage debates had run for decades before 1955. The Child Marriage Restraint (Amendment) Act of 1978, Act 2 of 1978, raised the ages to twenty-one for the bridegroom and eighteen for the bride, and the same amendment swept away the apparatus of guardianship in marriage, omitting the clause of Section 5 that had required a guardian’s consent for a bride under eighteen and repealing Section 6, which had listed the persons entitled to give that consent. Breach of the age condition has a peculiar consequence structure that traps many readers. The marriage is neither void under Section 11 nor voidable under Section 12, neither of which lists clause (iii) among the contraventions it sanctions; the breach is punishable under Section 18, and the bride married below fifteen acquired, as will be seen, a special right to repudiate the marriage and obtain a divorce. The dedicated child marriage statute, the Prohibition of Child Marriage Act of 2006, supplies its own annulment regime and is the owner of that subject in this series; within the Hindu Marriage Act the age condition operates principally as a penal and repudiation matter, and anyone who states that an underage Hindu marriage is automatically void has confused two statutes.

The fourth and fifth conditions, prohibited relationship and sapinda, are the conditions custom may displace, and the displacement must be the custom governing each of the parties, not one of them. The requirement that custom govern both is the working rule in the reported disputes: a marriage between parties within the prohibited degrees, valid by the custom of the husband’s community and unknown to the wife’s, fails the condition. The customs most often proved are those of communities in the south of India permitting marriage between cross-cousins and between a man and his sister’s daughter, relationships squarely within the statutory degrees and squarely within the custom that the Act preserves. The party who relies on a custom must plead and prove it, and the proof must satisfy the Section 3 definition: long, continuous and uniform observance, certainty, reasonableness and consistency with public policy. Where the condition is breached and no custom saves the marriage, the marriage is void under Section 11, like a bigamous marriage and unlike a marriage in breach of the capacity condition.

The five conditions therefore divide into three legal kinds. Breach of the monogamy, prohibited relationship or sapinda condition makes the marriage void, a nullity from the beginning. Breach of the capacity condition makes it voidable, valid until annulled. Breach of the age condition makes it neither, and sounds in punishment and in the bride’s right of repudiation. This tripartite structure is the single most examined feature of the Act’s entry law, and it is deliberate draftsmanship rather than accident: Parliament graded the conditions by the interest each protects, treating the existence of a living spouse and the forbidden degrees as matters of public order on which the parties’ acquiescence cannot confer validity, mental capacity as a matter principally between the parties, and age as a protective rule enforced by the criminal law and by an option given to the person the rule exists to protect.

The custom questions repay a closer look, because they are where the conditions meet the anthropology of the country the Act governs. The prohibited degrees in Section 3(g) and the sapinda limits in Section 3(f) describe the kinship of the northern, exogamous tradition, in which marriage within the gotra and within the counted generations is forbidden; large parts of the south observe the opposite preference, treating marriage between cross-cousins, and in some communities between a man and his elder sister’s daughter, as the preferred alliance that keeps property and kinship within the family. The Act holds both systems by the device of the custom exception, and the litigation it generates is of a predictable kind: a marriage within the degrees is attacked after it has failed, by a party who discovers the degrees when the relationship ends, and the defence is a custom that must be proved by instances, by the evidence of community elders and by the antiquity and uniformity the definition demands. The courts have insisted that the custom govern each party, so a marriage arranged across communities, valid in the custom of one and forbidden in the custom of the other, cannot be saved by the more permissive rule; and they have insisted equally that custom, once proved, is a complete answer, so that a marriage the text appears to forbid is valid in law because the community has always made it. The condition is therefore best understood as a default rule of exogamy with a community override, and the override, not the default, is what decides the disputed cases.

Why did the Marriage Laws (Amendment) Act 1976 matter so much?

The amendment of 1976, enacted on the recommendations in the fifty-ninth report of the Law Commission, was the most extensive revision the statute has undergone: it recast the mental capacity condition, added cruelty and desertion as grounds of divorce, and inserted divorce by mutual consent. No later amendment has changed the Act’s shape as much.

It also inserted the power to grant lesser relief on a divorce petition and shortened the waiting periods attached to decrees already obtained.

The fuller account of that amendment belongs to the sections on divorce below, where its provisions are taken one by one. Its place in the article at this point is to mark that the Act a reader meets in a bare text is a layered document: a 1955 core, a significant recasting in 1964 of the grounds that follow earlier decrees, the 1976 overhaul, the age revision of 1978, the epilepsy deletion of 1999, the procedural provisos of 2001 on the disposal of interim applications, the wife’s forum clause of 2003, and the leprosy deletion of 2019. A proposition about the Act is incomplete until the layer it comes from is identified, because the layers embody different theories of marriage: the fault theory of 1955, the limited breakdown element introduced in 1964, and the consent theory added in 1976, all three of which coexist in the text without reconciliation.

Ceremonies: the seventh step, and what makes the marriage complete

Section 7 provides that a Hindu marriage may be solemnised in accordance with the customary rites and ceremonies of either party to it, and that where such rites and ceremonies include the saptapadi, the taking of seven steps by the bridegroom and the bride jointly before the sacred fire, the marriage becomes complete and binding when the seventh step is taken. The provision is short and its consequences are large. The Act prescribes no form of its own. The ceremonies that make a marriage are the ceremonies of the parties’ communities, which may be elaborate or, where custom permits, minimal, and the statute’s only fixed point is the saptapadi rule, which resolves the recurring dispute about the moment at which a marriage becomes irrevocable in those rites that include the seven steps: not at the exchange of garlands, not at the tying of the sacred thread, but at the seventh step.

Two working rules follow. The first is that the party asserting a marriage must prove the ceremonies by which it was solemnised, and where the saptapadi forms part of the applicable rites, must prove its performance; convictions for bigamy have failed on exactly this footing, because a prosecution that cannot prove a valid first or second marriage cannot prove the offence built on it. Against that strictness stands the counterweight of the presumption from cohabitation: where a man and a woman have lived together for a long period as husband and wife, the law presumes in favour of marriage and against concubinage, and the burden on the party denying the marriage becomes heavy, a presumption the Supreme Court stated in Badri Prasad v Deputy Director of Consolidation, decided in 1978 and reported at (1978) 3 SCC 527, in the context of a union of some fifty years. The two rules meet in the contested cases: the presumption relieves the spouse who cannot produce the priests and witnesses of a ceremony decades old, while the prosecution for bigamy, which must prove its ceremonies to the criminal standard, receives no such assistance, so that the same union may be a marriage for maintenance and succession and unproved for the offence that Section 17 creates. Cognizance of that offence is itself restricted: the criminal procedure law permits a court to take cognizance of the offence of bigamy only on the complaint of a person aggrieved by it, which confines the prosecution in practice to the spouse and the close relations the provision names, and keeps the criminal sanction where the Act’s design places it, as the instrument of the party wronged rather than of public curiosity. The second is that custom governs here as it governs the conditions: a community whose rites do not include the saptapadi marries validly by its own rites, and Section 7(1) protects that variety by referring the question to the ceremonies of either party. The Act’s respect for form is thus the mirror of its severity on substance. It will dissolve a marriage for a single act of adultery proved to its satisfaction, and it will refuse to recognise a union of thirty years if the ceremonies that the parties’ own tradition requires were never performed.

Registration sits awkwardly beside this scheme, and the awkwardness is the point on which the brief for this article specifically directs attention, because the statutory position and the judicial direction pull in different directions. Section 8 empowers a state government to make rules for the registration of Hindu marriages, providing that the parties may have the particulars of their marriage entered in a Hindu Marriage Register kept for the purpose, and it does so expressly for the facility of proof of Hindu marriages. The same section then states that the validity of a marriage is in no way affected by omission to register it. Registration under the Act as enacted is therefore evidentiary machinery, optional in structure even where a state makes rules, and a marriage unregistered is as valid as a marriage registered if the ceremonies were performed.

One community within the Act’s application illustrates the scheme’s flexibility. Sikh marriages are within Section 2, and the Anand Karaj ceremony is a customary rite of the parties for the purposes of Section 7; the Anand Marriage Act of 1909 had long recognised that rite, and its amendment in 2012 supplied a dedicated registration mechanism for marriages solemnised by it, leaving validity, as this Act does, to rest on the ceremony and using registration for proof. The pattern is the Act’s pattern repeated at the margin: ceremony makes the marriage, the register proves it, and confusion between the two operations is the source of most error in the popular accounts.

In Seema v Ashwani Kumar, decided in 2006 and reported at (2006) 2 SCC 578, the Supreme Court, in proceedings in which the states and the National Commission for Women were heard, directed that marriages of all persons who are citizens of India belonging to various religions should be made compulsorily registrable in their respective states, and it set a timetable under which states were to notify procedures, invite objections and frame rules, specifying the officer who would register marriages and the particulars, including age and marital status, that the record should state. The Court returned to the subject in later orders in the same proceedings as compliance was reported state by state. The direction rests on the protective functions a register performs: proof of marriage for a spouse whose status is denied, a record of age against child marriage, a record of marital status against bigamy, and documentary support for claims to maintenance, custody and inheritance. The result is a dual position that a couple has to understand as a whole. Under the statute, registration is not a condition of validity and its absence invalidates nothing. Under the Court’s direction, carried into the rules of most states, registration is compulsory as an administrative obligation, and its absence forfeits the evidentiary presumption and the documentary ease that registration would have supplied, and may attract whatever consequence the state rules prescribe. A page that says registration is compulsory under the Hindu Marriage Act misstates the Act; a page that says registration does not matter misstates the practical law. Both errors are common, and the second is the more expensive one for the spouse, usually the wife, who later has to prove the marriage against denial.

The register’s function in litigation is evidentiary and should be understood exactly. An entry in the Hindu Marriage Register, and the certificate drawn from it, is strong documentary evidence that a marriage was solemnised between the persons named, and it shifts the practical burden in disputes over status: the spouse who denies the marriage must explain the entry away. It is not conclusive of validity. A registration entry cannot cure a marriage void for a living spouse or for prohibited degrees, and the absence of an entry cannot void a marriage whose ceremonies are proved, since Section 8(5) states the position in terms. In the states that framed compulsory registration rules under the Supreme Court’s direction, the obligation to register is administrative and is enforced, where it is enforced, by the consequences the rules prescribe and by the practical disabilities of the unregistered spouse, who must prove the marriage by witnesses to ceremonies performed years earlier, often after the officiants and elders have died. That asymmetry, between a ceremony the law treats as complete at the seventh step and a proof problem that grows with every year the register is left unsigned, is the lived tension between the Act’s sacramental form and its documentary machinery.

Marriages that were never valid, and marriages valid until annulled

Section 11 declares that any marriage solemnised after the commencement of the Act is null and void if it contravenes the monogamy condition, the prohibited relationship condition or the sapinda condition, and that it may, on a petition presented by either party to it, be so declared by a decree of nullity. The decree is declaratory in substance: a void marriage is no marriage in the eyes of the law from the moment of solemnisation, and either party may treat it as void, but the decree is what converts that legal position into a record on which status, remarriage and prosecution can safely proceed. The consequences reach beyond the parties. A second marriage contracted while a void first marriage subsists is itself exposed to nullity; a party to a void marriage who marries again does not commit bigamy by that act alone, since the first union was no marriage, though the party who knowingly goes through a void ceremony and then relies on its invalidity will find the courts unsympathetic in maintenance and estoppel arguments that the bare logic of nullity might suggest. The clean rule is that status follows the decree, and prudence follows the rule.

Section 12 governs the different category of the voidable marriage, which is valid for all purposes unless and until a decree of annulment avoids it. Four grounds support annulment. The marriage has not been consummated owing to the impotence of the respondent. The marriage is in contravention of the mental capacity condition in Section 5(ii). The consent of the petitioner, or of the petitioner’s guardian where a guardian’s consent was required under the law as it earlier stood, was obtained by force or by fraud as to the nature of the ceremony or as to any material fact or circumstance concerning the respondent. The respondent was, at the time of the marriage, pregnant by some person other than the petitioner. Each ground is then fenced by the limitations in Section 12(2), which are the part of the section petitioners lose on. A petition on force or fraud must be presented within one year after the force had ceased to operate or the fraud was discovered, and it fails if the petitioner lived with the other party as husband or wife with full consent after the force ceased or the fraud was discovered. A petition on the pregnancy ground requires the court to be satisfied that the petitioner was ignorant of the alleged fact at the time of the marriage, that proceedings were instituted within the time the section prescribes, and that marital intercourse with the petitioner’s consent did not take place after the fact was discovered. Annulment is therefore a remedy with a short fuse and a conduct bar, built on the theory that a party who knows of the defect and continues the marriage has elected to keep it.

Between the two sections sits Section 16, which performs the Act’s most quietly radical work. It provides that notwithstanding that a marriage is null and void under Section 11, any child of such marriage who would have been legitimate if the marriage had been valid shall be legitimate, whether the child was born before or after the commencement of the Marriage Laws (Amendment) Act of 1976, and whether or not a decree of nullity is granted and whether or not the marriage is held to be void otherwise than on a petition under the Act. The same legitimacy is extended to children of a voidable marriage annulled by decree, including children begotten or conceived before the decree and children born after it in the circumstances the section specifies. The protection is then bounded by Section 16(3): nothing in the section confers on such a child any rights in or to the property of any person other than the parents, where, but for the passing of the Act, the child would have been incapable of possessing or acquiring such rights by reason of not being legitimate. The child of a void marriage is therefore legitimate, inherits from the parents, and takes no rights against the wider family by virtue of the section: a settlement whose succession consequences within the joint family are traced in the article on the succession amendment, which owns that ground. Section 16 as it stands is itself the 1976 substitution; the original text had made the legitimacy of children of a void marriage depend on a decree of nullity having been passed, a dependency the amendment removed as indefensible, since it punished the child for the parents’ failure to litigate.

Nullity and money interact in a way the parties rarely foresee. A decree of nullity is a decree under the Act, and the court passing it may order permanent alimony and maintenance under Section 25, a position the Supreme Court recognised in Chand Dhawan v Jawaharlal Dhawan, decided in 1993, where it held that the power to award maintenance attaches to the decrees the Act authorises, including a decree of nullity, so that a spouse whose marriage is declared void is not by that declaration alone excluded from the Act’s financial protection. The contrast with the general maintenance jurisdiction is instructive: in Yamunabai Anantrao Adhav v Anantrao Shivram Adhav, decided in 1988, the Supreme Court held that a woman whose marriage is void because her husband’s first marriage was subsisting is not a legally wedded wife for the purpose of the maintenance provision of the criminal procedure code, a holding that leaves the second wife of a bigamous union dependent on the remedies she can obtain in the nullity proceeding itself and on the prosecution that Section 17 invites. The asymmetry is a deliberate product of the texts, and it is one of the points at which the Act’s treatment of the deceived second spouse remains open to the criticism that the law punishes her status rather than his deception.

The penal provisions complete the entry law. Section 17 has been noticed with the monogamy condition: a marriage between two Hindus solemnised while either has a spouse living is void, and the provisions of Sections 494 and 495 of the Indian Penal Code apply accordingly, so that the second ceremony is both a nullity and a crime. Section 18 attaches punishment to other breaches of the conditions. Contravention of the age condition is punishable, under clause (a) as substituted by the Prohibition of Child Marriage Act of 2006, with rigorous imprisonment which may extend to two years, or with fine which may extend to one lakh rupees, or with both, a substitution that aligned the marriage Act’s penalty with the child marriage statute’s own scheme. Contravention of the prohibited relationship and sapinda conditions, where the marriage is procured in breach of them, is punishable under clause (b) with simple imprisonment which may extend to one month, or with fine which may extend to one thousand rupees, or with both. The disparity between the two penalties is instructive: the age condition, whose breach leaves the marriage standing, carries the heavy sentence, while the kinship conditions, whose breach voids the marriage itself, carry a nominal one, the legislature having treated the nullity as the principal sanction for the one and the criminal law as the only sanction for the other. Punishment under the section is in addition to, and not in substitution for, the civil consequences the earlier sections describe, so that a party to a void bigamous marriage faces the Penal Code through Section 17 and a party to an underage marriage faces Section 18, while the marriage in the first case is nothing and in the second case subsists. The pattern is consistent with the grading of the conditions: the criminal law stands behind the conditions Parliament treated as public order, and leaves the voidable ground of incapacity to private election.

Restitution of conjugal rights: the remedy that asks the marriage to resume

Section 9 provides that when either the husband or the wife has, without reasonable excuse, withdrawn from the society of the other, the aggrieved party may apply by petition to the district court for restitution of conjugal rights, and the court, on being satisfied of the truth of the statements in the petition and that there is no legal ground why the application should not be granted, may decree restitution accordingly. An explanation added by the amendment of 1976 places the burden of proving reasonable excuse on the person who has withdrawn from the society of the petitioner. The decree orders nothing physical. It declares that the withdrawing spouse is to return to the society of the petitioner, and its enforcement under the Code of Civil Procedure proceeds by attachment of the judgment-debtor’s property rather than by imprisonment or compulsion of the person, a point the Supreme Court emphasised when the provision’s validity was challenged. What the decree does in practice is threefold. It formalises the court’s finding that the separation is without reasonable excuse, which bears on maintenance claims that depend on a spouse’s refusal to live with the other without sufficient reason. It starts a clock: non-restitution for one year after the decree is itself a ground of divorce under Section 13(1A), available to either party, including, as the Supreme Court held in Dharmendra Kumar v Usha Kumar in 1977, the party whose own failure to comply produced the ground. And it functions, in the experience of the family courts, as a tactical instrument, filed as often to create a record or to answer a maintenance or criminal proceeding as to recover a companionship that a decree cannot manufacture. The remedy’s defenders and its critics both begin from that record, and reach opposite conclusions about what it shows.

The working law of the section is contained in the phrase “without reasonable excuse” and in the allocation of the burden the 1976 explanation enacted. The petitioner proves the withdrawal from his or her society; the respondent who resists must prove that the withdrawal was with reasonable excuse, and the courts have treated the recognised matrimonial wrongs as excuses in themselves: a husband’s cruelty, his keeping of another woman in the house, his persistent neglect or his expulsion of the wife from the home each answer the petition, because a spouse cannot be in default for refusing a society that the petitioner has made unreasonable to share. A decree obtained has consequences beyond its own terms. A wife who refuses without sufficient reason to live with her husband after a decree in his favour imperils her claim to maintenance under the general maintenance law, which denies maintenance to a wife who refuses to live with her husband without sufficient reason, and the decree is the court’s finding that no sufficient reason exists; the same finding arms the husband’s defence in proceedings under the Hindu Adoptions and Maintenance Act, where a wife’s entitlement to live separately depends on the grounds that statute lists. This is the mechanism by which a decree that orders nothing physical nonetheless redistributes the parties’ legal positions across every connected proceeding, and it is why restitution petitions are so often the opening move of a husband who expects to be sued for maintenance, rather than the plea of a spouse seeking reunion. The family courts know the pattern and say so in their judgments; the provision’s reconciling purpose survives in the section’s text and in the reconciliation duty of Section 23, and its adversarial use survives in the decided cases, and an honest account of the remedy has to carry both.

Why is restitution of conjugal rights constitutionally controversial?

The constitutional objection is that a decree ordering one adult to resume cohabitation with another invades the liberty, privacy and decisional autonomy that Article 21 of the Constitution protects, and does so in a form whose burden falls overwhelmingly on wives.

The countervailing view is that the decree compels no act of the body, is enforceable only against property, serves the social purpose of holding a marriage open to reconciliation, and is available to either spouse on equal terms. Both positions have been held by constitutional courts, and the question is once again before the Supreme Court.

The litigation has a shape worth setting out in full, because it is one of the few places where the Act’s sacramental inheritance was tested directly against the Constitution. In T. Sareetha v T. Venkata Subbaiah, decided in 1983 and reported at AIR 1983 AP 356, a single judge of the Andhra Pradesh High Court, Choudary J., held Section 9 unconstitutional and void. The judgment reasoned that a decree of restitution is a coercive order directed at the most intimate choices a person makes, that it treats the wife’s body as subject to a husband’s claim enforceable by the state, and that the remedy offends the right to privacy and human dignity within Article 21, beside failing the equality guarantee in its real-world incidence. Within a year the Delhi High Court held the opposite in Harvinder Kaur v Harmander Singh, reported at AIR 1984 Del 66, where Rohatgi J. upheld the section, reading the decree as an inducement to cohabitation rather than a compulsion to sexual relations, and observing pointedly that the provision also operates as a route by which a spouse may in time obtain a divorce. The conflict went to the Supreme Court, which in Saroj Rani v Sudarshan Kumar Chadha, decided in 1984 and reported at (1984) 4 SCC 90, upheld the section and disapproved the Andhra Pradesh view. The Court held that the object of the decree is to bring about cohabitation between estranged parties, that it serves a social purpose of preserving the marriage, that conjugal rights in the section’s sense comprehend consortium and companionship rather than enforced sexual relations, and that the safeguards of the execution law, under which the decree is enforced against property and not by detention of the person, answer the charge of barbarity.

That settlement has not held in argument, whatever its formal authority. The constitutional law around it has moved: the recognition of privacy as a fundamental right with decisional autonomy at its core, and the decriminalisation of adultery on reasoning that treated the marriage relation as one between equals rather than as a hierarchy, have given the Sareetha argument a doctrinal foundation it lacked in 1983. A fresh challenge to the restitution provisions across the marriage statutes, brought in Ojaswa Pathak v Union of India in 2019, attacks Section 9 of this Act together with its analogues, and the matter remains pending before the Supreme Court; it is described here as pending, and no outcome is asserted. The stance this article takes on the merits is the one the record supports. The decree’s formal enforcement is genuinely weak, and the Saroj Rani account of it is accurate as far as it goes; but the decree’s practical uses, as a maintenance defence and as the first move in a divorce campaign, are the strongest part of the case against it, and they are uses the 1984 judgment did not examine. A reader should hold both: the provision stands valid, its constitutional foundations are under renewed and serious question, and its working life in the courts is more adversarial than its reconciling theory suggests.

Judicial separation: the decree that suspends cohabitation without ending the marriage

Section 10 entitles either party to a marriage, whether solemnised before or after the commencement of the Act, to present a petition for a decree for judicial separation on any of the grounds specified in Section 13(1), and, in the case of a wife, on any of the grounds specified in Section 13(2). The decree’s effect, stated in the section itself, is that it shall no longer be obligatory for the petitioner to cohabit with the respondent. The marriage subsists in every other respect: neither party may remarry, the status and the obligations that do not depend on cohabitation continue, and the court may, on the application by petition of either party and on being satisfied of the truth of the statements in it, rescind the decree if it considers it just and reasonable to do so. Judicial separation is the remedy for a spouse who needs the law’s protection against compelled cohabitation, and needs time or leverage, but who does not want, or cannot yet prove the case for, a dissolution.

The section’s present form is the 1976 settlement. As enacted in 1955, judicial separation had its own shorter list of grounds, and cruelty and desertion stood on it; divorce, by contrast, demanded graver and more narrowly drawn grounds, so that the two remedies formed a ladder on which the lesser relief was available for the lesser wrong. The amendment of 1976 aligned the grounds: the same facts that found a divorce found a judicial separation, and the choice between them became a choice about outcome rather than about proof. A petitioner who proves cruelty may take a decree that ends the marriage or a decree that suspends it, and the court’s power under Section 13A, inserted by the same amendment, to grant the lesser relief of judicial separation on a petition for divorce, where it considers it just to do so having regard to the circumstances, completes the alignment. One consequence of the alignment is strategic and well understood in practice: a decree of judicial separation starts the same one-year clock as a decree of restitution, so that separation maintained for a year after the decree ripens into a ground of divorce under Section 13(1A) for either party. The lesser remedy is thus also, in the hands of a party who wants the marriage ended, the first instalment of the greater one, a feature the Law Commission accepted as the point at which the statute admits a measured element of breakdown into a fault code.

The routes out of a marriage are best seen together before the divorce grounds are examined in detail, because the choice among them is the decision every reader of this statute actually faces.

Why a party chooses judicial separation over divorce, when the facts support either, is a question with practical rather than doctrinal answers. The petitioner may have a religious or conscientious objection to dissolving the marriage while needing the law’s protection against cohabitation and its incidents. The petitioner may wish to preserve the status of spouse for purposes that depend on it, in maintenance negotiations, in claims on the other’s employment benefits and nominations, and in the expectations of a joint family in which a divorced daughter-in-law and a separated one occupy different positions. The petitioner may not yet be able to prove the case to the standard a divorce requires and may prefer to fix the separation by decree while the evidence matures. And the petitioner may simply want time: the decree can be rescinded if the parties reconcile, which a divorce decree cannot. The price of the lesser remedy is that the parties remain married, with everything that follows: no remarriage, continuing exposure to the other’s claims, and succession rights between them that a divorce would have ended. The choice is therefore a real one, and the statute’s alignment of the grounds in 1976 was designed to make it a free one, turning on what the petitioner wants the decree to do rather than on what the petitioner can prove.

Which route and how long: the ways a Hindu marriage ends

Route	Legal basis	Who may seek it	What must be proved	Minimum time	Forum	Effect on maintenance and custody
Nullity of a void marriage	Section 11	Either party to the marriage	Breach at solemnisation of the monogamy, prohibited relationship or sapinda condition in Section 5	No waiting period; the marriage is void from the start	District court (family court)	Marriage treated as never having existed; children of the marriage are legitimate under Section 16; maintenance and expenses may be ordered within the proceeding under Sections 24 and 25
Annulment of a voidable marriage	Section 12	The party aggrieved by the ground relied on	A ground in Section 12(1), within the time limits and conduct bars of Section 12(2)	One year from cessation of force or discovery of fraud on the force and fraud grounds	District court (family court)	Marriage valid until annulled; children legitimate under Section 16; interim and permanent maintenance available within the proceeding
Judicial separation	Section 10	Either party; a wife also on the Section 13(2) grounds	Any ground in Section 13(1), proved to the standard Section 23 requires	None beyond the period built into the ground itself, such as two years for desertion	District court (family court)	Marriage subsists and remarriage is barred; maintenance and custody orders available; decree may be rescinded on application
Contested divorce	Sections 13(1), 13(1A) and 13(2)	Either party; a wife alone on the Section 13(2) grounds	The ground alleged, proved on the petitioner’s own case without taking advantage of the petitioner’s own wrong	Two years for desertion; one year of non-resumption after a separation or restitution decree; first-year bar in Section 14	District court (family court)	Marriage dissolved; permanent alimony under Section 25 and custody under Section 26 may be ordered in the decree
Divorce by mutual consent	Section 13B	Both parties jointly	Living separately for one year or more, inability to live together, and mutual agreement to dissolve	Second motion not earlier than six months and not later than eighteen months after the petition; the six months may be waived	District court (family court)	Marriage dissolved on the decree; alimony, custody and property terms as the parties settle and the court records
Dissolution for irretrievable breakdown	Article 142 of the Constitution; no provision in the Act	A party to proceedings reaching the Supreme Court; no one may claim it as a matter of right	Complete and irretrievable breakdown of the marriage, to the satisfaction of the Supreme Court on the factors it has laid down	No statutory period; long separation, including six years or more, weighs heavily	Supreme Court only	Dissolution with such directions on alimony and connected proceedings as the Court makes to do complete justice in the matter

The final row is the artifact’s payload and the article’s central fact: the route that answers the commonest situation in matrimonial failure, a marriage that has simply ended, is not in the statute at all.

The grounds of divorce in Section 13: the fault code

Section 13(1) states that any marriage solemnised, whether before or after the commencement of the Act, may, on a petition presented by either the husband or the wife, be dissolved by a decree of divorce on the grounds it lists. The list in its present form is the product of three legislatures, and it cannot be understood without its layers. The Act of 1955 enacted nine grounds resting on the fault theory, with two features that strike a modern reader immediately: adultery had to be proved as a course of conduct, the respondent “living in adultery”, and cruelty and desertion were absent altogether, being grounds only for judicial separation. The reasoning of the period was that a wrong that might be transient should suspend a marriage rather than end it, and that divorce should be reserved for the wrongs and disabilities that made married life impossible or dangerous. The Hindu Marriage (Amendment) Act of 1964, Act 44 of 1964, recast the last two of the original grounds, which had allowed divorce where cohabitation had not resumed after decrees of judicial separation or restitution, into a new sub-section, Section 13(1A), available to either party to the earlier proceeding and carrying a two-year period. The amendment of 1976 then rebuilt the list: it substituted the adultery clause so that a single act of voluntary sexual intercourse with a person other than the spouse suffices; it inserted cruelty and desertion as grounds of divorce in clauses (ia) and (ib); it recast the mental condition clause in the functional language already noticed in Section 5; it removed the minimum periods that had attached to the disease grounds; and it reduced the Section 13(1A) period from two years to one.

The grounds in their present form, taken in the statute’s order, are these. Either party may petition on the ground that the other party has, after the solemnisation of the marriage, had voluntary sexual intercourse with any person other than his or her spouse; that the other party has treated the petitioner with cruelty after the solemnisation of the marriage; that the other party has deserted the petitioner for a continuous period of not less than two years immediately preceding the presentation of the petition; that the other party has ceased to be a Hindu by conversion to another religion; that the other party has been incurably of unsound mind, or has been suffering continuously or intermittently from mental disorder of such a kind and to such an extent that the petitioner cannot reasonably be expected to live with the other party; that the other party has been suffering from venereal disease in a communicable form; that the other party has renounced the world by entering any religious order; or that the other party has not been heard of as being alive for a period of seven years or more by those persons who would naturally have heard of the party if the party had been alive. An explanation to the mental condition clause defines mental disorder to mean mental illness, arrested or incomplete development of mind, psychopathic disorder or any other disorder or disability of mind, and includes schizophrenia; psychopathic disorder is itself defined as a persistent disorder or disability of mind, whether or not including subnormality of intelligence, which results in abnormally aggressive or seriously irresponsible conduct and requires or is susceptible to medical treatment. Desertion is likewise explained: the expression includes desertion without reasonable cause and without the consent or against the wish of the petitioner, and includes the wilful neglect of the petitioner by the other party, and its grammatical variations and cognate expressions are to be construed accordingly.

One clause of the original list no longer appears, and its disappearance is itself a piece of the Act’s history that the brief for this article directs be stated accurately, because out-of-date texts continue to reproduce it. The ground of virulent and incurable leprosy stood in Section 13(1)(iv) from 1955, softened in 1976 by the removal of its three-year period, and it was omitted altogether by the Personal Laws (Amendment) Act of 2019, Act 6 of 2019, which received the assent of the President on 21 February 2019 and amended five personal law statutes to remove leprosy as a ground of divorce or of denial of maintenance, on the stated ground that the disease is curable, that the medical position had changed, and that retaining a curable condition as a matrimonial disability discriminated against persons affected by it. A petition that pleads leprosy as a ground pleads a provision Parliament has repealed; the error survives on websites that copy unrepealed texts, and it is one of the markers by which stale accounts of this Act may be known.

Adultery and the two grounds inserted in 1976 carry the great bulk of contested litigation, and each has a working jurisprudence that a pillar article must summarise without pretending to exhaust. Adultery after 1976 is a single voluntary act, proved, like all matrimonial offences, on the preponderance of probabilities rather than to the criminal standard, and usually by circumstantial evidence, since direct evidence is rare; the decriminalisation of adultery by the Constitution Bench in Joseph Shine v Union of India in 2018, which struck down the penal provision, expressly left its character as a matrimonial wrong untouched, and it remains a ground of divorce and a consideration in maintenance. Cruelty is the ground on which the case law is deepest. Desertion is the ground on which it is most technical: the petitioner must prove both the fact of separation and the intention to bring cohabitation permanently to an end, the animus deserendi of the authorities, together with the absence of reasonable cause and of consent, and must prove them across the whole of the two-year period immediately before the petition, so that a separation that began in desertion may cease to be desertion if the deserter makes a genuine offer to return, and a separation by consent is never desertion at all. Constructive desertion completes the doctrine: the spouse whose conduct drives the other from the home is the deserter in law, however the physical departure is arranged, which is why the question in a desertion case is never simply who left.

The remaining grounds are narrower in use but each carries a settled construction. Conversion requires that the respondent has ceased to be a Hindu by conversion to another religion; the petitioner’s own religion is irrelevant, the ground is complete on proof of a genuine conversion, and, as the bigamy cases below establish, the marriage is not dissolved by the conversion itself but only by the decree the petitioner obtains. Renunciation requires that the respondent has renounced the world by entering a religious order; the courts require a formal entry into an established order, with the civil death that orthodox doctrine attaches to sanyasa, and a mere change of dress, title or devotional habit does not suffice. Presumed death requires that the respondent has not been heard of as alive for seven years or more by those who would naturally have heard of the respondent if alive, a formulation that works with the presumption in Section 108 of the Indian Evidence Act, and the decree that follows is a decree of divorce: it dissolves the petitioner’s marriage so that the petitioner may remarry, and it does not unwind the dissolution if the missing spouse returns, which is why the ground is taken in practice where a second union is intended. Venereal disease must be in a communicable form; the statute does not require that the disease be of long standing since 1976 removed the durational element, and it does not require the petitioner to prove where the disease was contracted, the ground being the condition and its communicability rather than the respondent’s fault in acquiring it.

The mental condition ground is the most heavily evidenced in the list, because it asks the court to dissolve a marriage on medical material while protecting a respondent whose illness may prevent instructions being given. The statute’s two limbs run together in practice: incurable unsoundness of mind, and mental disorder of such a kind and to such an extent that the petitioner cannot reasonably be expected to live with the respondent, the second limb being the working test. In Ram Narain Gupta v Rameshwari Gupta, decided in 1988, the Supreme Court held that the statutory language requires more than eccentricity, neurosis or a difficult temperament: the disorder must be of the kind or extent that makes married life with the respondent unreasonable to expect, judged on the medical evidence and the conduct it explains. In Sharda v Dharmpal, decided in 2003 and reported at (2003) 4 SCC 493, the Court held that a matrimonial court has power to direct a party to undergo medical examination where a ground of this character is pleaded, and that a party who refuses without cause may face an adverse inference, while cautioning that the power is not to be exercised as a matter of routine or as an instrument of humiliation. The balance those two cases strike is the ground’s working law: the petitioner’s case must be proved by evidence of the disorder and its effects, not by the label of a diagnosis, and a diagnosis of schizophrenia, which the explanation expressly names, is no more a decree in itself than any other named condition.

How do the courts approach mental cruelty as a ground of divorce?

Mental cruelty has no statutory definition, and the courts have refused to give it one, holding instead that conduct must be assessed in the setting of the parties’ lives: the question is whether the behaviour complained of is of such a character that the petitioner cannot reasonably be expected to live with the respondent.

In V. Bhagat v D. Bhagat, decided in 1994 and reported at (1994) 1 SCC 337, the Supreme Court held that mental cruelty is conduct that inflicts such mental pain and suffering that the parties cannot be expected to live together, a state of mind and feeling rather than a list of acts.

The working propositions are stable enough to state. Cruelty is not the ordinary wear and tear of married life, and the courts repeat that the institution would be unworkable if every quarrel, coldness or disappointment founded a decree. Sustained humiliation, false and reckless allegations of unchastity or of mental illness, unfounded criminal complaints pursued to harassment, persistent refusal of marital relations without cause, and conduct that destroys the petitioner’s standing in family and profession have all been held capable of amounting to cruelty, with the emphasis on pattern and effect rather than on any single incident. Physical cruelty requires no such elaboration and never has. The ground interacts with the breakdown question in a way the courts acknowledge: where a marriage has collapsed and the parties have litigated against each other for years, the litigation itself generates the allegations from which cruelty is then found, so that the fault code, applied to a dead marriage, manufactures the fault it requires. That circularity is the strongest practical argument in the breakdown debate, and it is recorded here as the courts themselves have recorded it, including in the Supreme Court’s observation in K. Srinivas Rao v D. A. Deepa in 2013 that the filing of false criminal complaints can itself amount to mental cruelty entitling the other spouse to dissolution.

Section 13(1A) has been noticed already: either party to a marriage may petition for divorce on the ground that there has been no resumption of cohabitation for one year or more after the passing of a decree for judicial separation in a proceeding to which they were parties, or that there has been no restitution of conjugal rights for one year or more after the passing of a decree for restitution. Its importance is structural. It is the first place in the Act where the passage of time apart, without any fresh fault, dissolves a marriage, and it is the thin end of the breakdown principle inside a fault statute; its limitation to parties who have already obtained a decree is the measure of how far Parliament in 1964 was prepared to go, and no further. The wife’s additional grounds in Section 13(2) are of a different character. A wife may petition on the ground that, in the case of a marriage solemnised before the commencement of the Act, the husband had married again before that commencement, or another wife of the husband married before that commencement was alive at the time of her own marriage, a clause that addressed the pre-Act polygamous marriages the Act could not retrospectively void; that the husband has, since the solemnisation of the marriage, been guilty of rape, sodomy or bestiality; that a decree or order awarding her maintenance has been passed in her favour under the Hindu Adoptions and Maintenance Act or in a proceeding under the general maintenance provision of the criminal procedure code, and that cohabitation has not been resumed for one year or upwards since that decree or order notwithstanding that she was living apart; or that her marriage, whether consummated or not, was solemnised before she attained the age of fifteen years and she has repudiated the marriage after attaining that age but before attaining the age of eighteen years. The four grounds repay attention as a group, because they record the wrongs the general list does not capture. The first is transitional and closed: it protected the wives of pre-Act polygamous marriages, whom the Act refused to treat as parties to void unions but would not leave remediless. The second imports the gravest sexual offences into matrimonial relief, so that a wife need not wait on a prosecution to escape the marriage in which they were committed. The third converts an existing maintenance decree into an exit where the husband, ordered to maintain a wife he will not live with, allows a year to pass without resuming cohabitation: it is in substance a breakdown ground confined to marriages already reduced to a maintenance order, and its drafting, which requires the decree or order to have been passed in proceedings to which the husband was a party and the wife to have been living apart when it was passed, has generated its own case law on what counts as resumption. The fourth is of a different order again, and is taken separately below. The third and fourth of these grounds were inserted by the amendment of 1976; the fourth, the repudiation ground, is the Act’s own answer to the child bride, an option of puberty exercised through divorce rather than annulment, and it applies whether the marriage was solemnised before or after the 1976 amendment commenced.

Two procedural provisions shape every contested petition. Section 14 bars a petition for divorce presented within one year from the date of the marriage, subject to the court’s power to allow an earlier petition on the ground of exceptional hardship suffered by the petitioner or exceptional depravity on the part of the respondent, and a decree obtained by misrepresentation or concealment in that leave application may be made conditional or the petition dismissed without prejudice. Section 23 then states the conditions on which decrees may be granted, and it converts the fault theory into rules of decision.

What must a court be satisfied of before granting a matrimonial decree?

Section 23 requires the court, before decreeing relief, to be satisfied that the ground exists and that the petitioner is not taking advantage of his or her own wrong, that adultery or cruelty relied on has not been condoned, that there is no collusion between the parties, and that the petition has not been unreasonably delayed.

Relief is refused to a petitioner who fails these conditions, however clearly the ground itself is proved.

The section’s companion provisions complete the procedure. Section 23(2) obliges the court, before granting relief, in the first instance to make every endeavour to bring about a reconciliation between the parties, except where the relief is sought on grounds that reconciliation cannot sensibly address, namely conversion, unsoundness of mind or mental disorder, venereal disease, renunciation of the world or presumed death; and Section 23(3) equips the court to adjourn the proceedings for a period not exceeding fifteen days and refer the matter to a person named by the parties or by the court for an attempt at reconciliation, with the report to be taken into account. Section 23A, also a creature of 1976, allows a respondent not merely to oppose the petition on the ground of the petitioner’s adultery, cruelty or desertion but to make a counter-claim for any relief under the Act on that ground, and if the petitioner’s wrong is proved the court may give the respondent the relief the respondent could have obtained by petitioning. Section 21B directs that a petition be tried, as far as practicable consistently with the interests of justice, on a day-to-day basis once trial begins, and that every petition be disposed of within six months from the date of service of notice on the respondent, and every appeal within three, directions honoured in the family courts with a fidelity the pendency figures contradict but which state the legislature’s intent about the tempo of matrimonial litigation. The general information in this article is not a substitute for advice on an individual matter, and matrimonial limitation periods, including the one-year bars in Sections 12 and 14 and the eighteen-month limit in Section 13B, are unforgiving of delay. Section 28 makes decrees appealable as decrees of original civil jurisdiction, and orders under Sections 25 and 26 appealable on law and fact, while no appeal lies on costs alone, and an order under Section 24, being interim, is not appealable under the section at all.

Two of the Section 23 conditions do work that surprises parties who have proved their ground. The bar on taking advantage of one’s own wrong means that a petitioner whose own adultery or cruelty provoked the respondent’s conduct may be refused a decree on the very ground that conduct supplies, the doctrine that produced, in the desertion cases, the rule that the spouse who drives the other away cannot petition on the departure. Condonation is the second: a matrimonial offence forgiven, with the parties reinstated to the position they occupied before it, cannot afterwards be revived as a ground, although fresh misconduct of a cognate kind revives the condoned offence with it, so that a single act after forgiveness restores the whole history to the petition. Both doctrines express the fault theory’s internal morality, that the court dissolves marriages for the innocent against the guilty, and both are the first casualties of the breakdown critique, since in a marriage that has failed without a villain neither doctrine has anything sensible to operate on. Their practical importance is undiminished: more contested petitions fail on Section 23 than fail on proof of the ground, and the section is the reason a matrimonial petition is drafted backwards, from the petitioner’s own conduct outward, rather than forwards from the respondent’s.

The Act’s machinery: petitions, transfer and the shape of a proceeding

A pillar article owes the reader the procedural skeleton on which the remedies hang, because in matrimonial litigation the forum and the sequence decide as much as the grounds. Section 19 has been noticed for its grounds of jurisdiction, including the wife’s residence clause of 2003. Section 20 requires that every petition state distinctly the nature of the case, the facts on which the claim is founded and, where relief is sought on a ground of the other party’s conduct, the particulars of that conduct, and it requires verification in the manner of a plaint; the pleading discipline matters because matrimonial grounds are pleaded in the alternative with great freedom and tested with corresponding severity, and a decree founded on a case not pleaded is vulnerable in appeal. The papers in such a case run from the marriage certificate and the record of ceremonies to income statements, bank records and the correspondence of the separation, and a party assembling the bundle will repeatedly need to open and read the Word documents and spreadsheets within it through an office file viewer before the papers are settled and filed. Section 21 applies the Code of Civil Procedure, subject to the Act and the rules of the High Court, to all proceedings under the Act. Section 21A, another insertion of 1976, deals with the recurring collision of a husband’s petition in one district and a wife’s in another, by providing for transfer and joint trial so that the same marriage is not litigated to inconsistent decrees in two courts. Section 21C protects documents from exclusion merely on the ground that they are not duly stamped or registered. The reconciliation duty in Section 23(2) and (3), the counter-claim in Section 23A and the day-to-day trial direction in Section 21B have been set out with the grounds, where they operate. Two features deserve separate emphasis. The first is that the Act’s proceedings are, by Section 22, held in camera and shielded from publication, a protection the parties may invoke and the court may impose, so that the record of a marriage’s failure is not public property in the way of ordinary civil litigation. The second is the practical reality that a contested petition travels through interim maintenance under Section 24, reconciliation attempts, evidence, argument and decree over a period measured in years in most districts, against which the statutory aspiration of disposal within six months stands as a direction of intent rather than a description. Anyone planning litigation under this Act plans on that timescale, and takes advice before a petition is filed, since the choice of ground, forum and first move is frequently irreversible in its consequences.

Divorce by mutual consent: Section 13B and the six months

The amendment of 1976 inserted Section 13B, and with it a second theory of divorce entered the Act beside the fault theory: dissolution by agreement. The section provides that a petition for dissolution of marriage by a decree of divorce may be presented to the district court by both the parties to a marriage together, whether the marriage was solemnised before or after the commencement of the amendment, on the ground that they have been living separately for a period of one year or more, that they have not been able to live together and that they have mutually agreed that the marriage should be dissolved. The second sub-section then prescribes the sequence that has made the provision the most searched text in Indian family law: on the motion of both the parties, made not earlier than six months after the date of the presentation of the petition and not later than eighteen months after that date, if the petition is not withdrawn in the meantime, the court shall, on being satisfied, after hearing the parties and after making such inquiry as it thinks fit, that a marriage has been solemnised and that the averments in the petition are true, pass a decree of divorce declaring the marriage to be dissolved with effect from the date of the decree.

Four working propositions govern the provision. Living separately means living not as husband and wife; the Supreme Court held in Sureshta Devi v Om Prakash, decided in 1991 and reported at (1991) 2 SCC 25, that the parties need not be living under different roofs, since a couple may share a residence and yet have ceased to live as spouses, and the same case settled that the mutual consent must subsist until the decree is passed, so that either party may withdraw consent before the second motion is heard and the court cannot proceed on consent that has been retracted. The one-year separation under sub-section (1) is a condition of the petition’s maintainability and is not dispensable by the trial court. The six-to-eighteen-month window under sub-section (2) is procedural in purpose, an interval for reflection built into the statute, and its character was the question in the waiver litigation. And the decree, when passed, dissolves the marriage from its own date, with the parties’ settlements on alimony, custody and property ordinarily recorded in the petition and given effect through the court’s orders under Sections 25 and 26.

In Amardeep Singh v Harveen Kaur, decided in 2017 and reported at (2017) 8 SCC 746, a two-judge bench of the Supreme Court, Adarsh Kumar Goel and Uday Umesh Lalit JJ., held that the period in Section 13B(2) is directory and not mandatory, and that the court dealing with the matter may waive it where the statutory conditions for waiver are satisfied. The conditions the judgment states are that the period of six months, in addition to the period of one year of separation under Section 13B(1), is already over before the first motion itself; that all efforts at mediation and conciliation, including under the reconciliation duty of the Act and the Family Courts Act, have failed and there is no likelihood of success by any further effort; that the parties have genuinely settled their differences, including alimony, custody of children and any other pending issues between them; and that the waiting period will only prolong their agony. The waiver application may be filed one week after the first motion, giving reasons. The decision displaced a line of earlier rulings in which parties had been driven to the Supreme Court itself to obtain, under its constitutional power, the waiver no lower court would give, and it converted an exceptional constitutional indulgence into a working discretion of the family courts. It did not touch the one-year separation, which remains mandatory, and later benches have been careful to keep the two periods distinct: the year is the substance of the ground, the six months are the statute’s pause for reflection, and only the pause is waivable.

The conduct of a mutual consent case has acquired its own settled practice, which the text only partly reveals. The petition is presented jointly, the parties’ statements are recorded at the first motion, and the court satisfies itself that the consent is free and that the terms recorded, on permanent alimony, the custody and upbringing of children, and the division or retention of property and effects, are the parties’ own. Where a party is residing abroad or cannot attend, the judgment in Amardeep Singh itself contemplated the use of video conferencing and representation through close family members holding authority, so that the geography of a separated couple’s lives does not defeat a consent both parties have genuinely given. Between the motions the petition may be withdrawn by either party, and the second motion is the point of no return: consent withdrawn before it defeats the petition entirely, however long the separation and however complete the settlement, which is the leverage that makes the six months something more than a waiting room. Where the parties do not move the second motion within eighteen months, the petition fails and the parties are remitted to a fresh petition or to the fault grounds, a lapse that occurs in practice where a reconciliation is attempted and fails slowly. Practitioners therefore treat the timetable as part of the settlement itself, fixing dates for the motions alongside the financial terms, and the court’s inquiry at the second motion, directed to the truth of the averments and the subsistence of consent, is ordinarily short where the record of the first motion is complete.

The provision’s practical significance exceeds its doctrinal novelty. Mutual consent is the route by which most Hindu divorces that conclude at all are concluded, because it substitutes negotiation for proof and fixes the terms of separation in advance; its weaknesses are the mirror of its strengths, since consent can be withdrawn tactically at the second motion, leaving the other party to begin again on a fault ground years into the separation, and since the quality of the settlement a weaker spouse accepts under the pressure of delay is a standing concern of the family courts and of the women’s organisations whose arguments recur in the breakdown debate. The statute gives the court no general power to test a mutual consent settlement for fairness beyond satisfying itself that the consent is genuine and the averments true, a thin control that both sides of the reform argument cite.

Two later developments complete the provision’s working law. The first is the working-out of Amardeep Singh in the courts below it. In Amit Kumar v Suman Beniwal, decided in 2021, the Supreme Court confirmed that the factors stated in Amardeep Singh are the matters the court seized of the petition is to weigh, and that the discretion to waive the six months belongs to that court, so that parties who satisfy the conditions need not travel to the Supreme Court for a relief the family court can give; benches since have treated the four conditions as guides to a discretion rather than as a checklist whose every item must be proved like a ground of divorce. The second is the use of the constitutional power to close the whole controversy at once. Mutual consent petitions in practice travel with connected proceedings, maintenance claims, custody disputes and criminal complaints arising out of the same marriage, and a settlement that dissolves the marriage while leaving the satellite litigation alive settles little. The Constitution Bench in Shilpa Sailesh, treated in the next section for its breakdown holding, answered the connected question as well: the Supreme Court, acting under Article 142 in a matter before it, may grant the divorce by mutual consent dispensing with the waiting period and the family court procedure, and may dispose of the connected proceedings between the parties so that the settlement is complete. The contrast between the two developments is the contrast that runs through this Act’s modern history: the ordinary courts received, by interpretation, a measured discretion over a procedural interval; the extraordinary power to override the statute’s sequence altogether remained where the Constitution put it, in one court.

Conversion and the second marriage: the bigamy line of cases

Section 13(1)(ii) gives a spouse a ground of divorce where the other party has ceased to be a Hindu by conversion to another religion, and the ground has a feature petitioners misunderstand: conversion does not itself dissolve the marriage. The converted spouse remains married under the Act until a decree dissolves the marriage, and the unconverted spouse must petition. From that simple proposition the most consequential personal law litigation of the 1990s followed. A practice had grown in which a Hindu husband, wishing to marry again without divorcing, converted to Islam and contracted a second marriage under the personal law he had adopted, claiming that the conversion had carried him out of the monogamous statute. In Sarla Mudgal v Union of India, decided in 1995 and reported at (1995) 3 SCC 635, the Supreme Court held that a marriage solemnised under the Act continues until it is dissolved under the Act, that conversion does not dissolve it, that a second marriage contracted during its subsistence is void, and that the husband who contracts it is liable to prosecution for bigamy under Section 494 of the Indian Penal Code. The Court accompanied the holding with observations on the desirability of a uniform civil code, which drew the case a political attention beyond its ratio.

In Lily Thomas v Union of India, decided in 2000 and reported at (2000) 6 SCC 224, the Court was asked in review and writ proceedings to reconsider Sarla Mudgal, on the footing among others that it violated the freedom of religion. The Court reaffirmed the holding: the second marriage of a Hindu husband after conversion, contracted without dissolution of the first, is invalid and the offence of bigamy is committed, and it clarified that Sarla Mudgal had issued no direction to the government to enact a uniform civil code, the observations on that subject being an expression of the desirability of reform rather than a mandate. The Law Commission’s two hundred and twenty-seventh report, in 2009, recommended giving the settled position statutory form by inserting express provisions into the marriage statutes, a recommendation that illustrates the series’ recurring pattern of a rule that is law in the courts awaiting a text in the statute. For the purposes of this Act the position is closed and should be stated flatly: conversion changes the convert’s religion; it does not change the convert’s marital status; a second marriage before a decree of divorce is void and punishable; and the first spouse’s remedies under Sections 11, 13 and 17 are unaffected by the ceremony’s validity under the law the convert has adopted.

The missing ground: irretrievable breakdown of marriage

Against every provision so far described stands the ground that is not there. A marriage may have ended in fact for a decade, the parties living in different cities or different countries, communicating only through counsel, with no allegation either wishes to prove and none that would survive Section 23, and no court in India except the Supreme Court may dissolve it on that ground. The history of the attempt to supply the omission is long, documented and unsuccessful. The Law Commission considered breakdown incidentally in its fifty-ninth report of 1974 and made no specific recommendation. In its seventy-first report, in 1978, the Commission recommended that irretrievable breakdown of marriage be added as a ground of divorce, reasoning that restricting divorce to a matrimonial offence or disability causes injustice where neither party is at fault, or the fault is of a kind the parties do not wish to divulge, and that there is hardly any utility in maintaining a marriage as a facade when the emotional and other bonds which are of the essence of marriage have disappeared. A bill introduced in 1981 to give effect to that recommendation did not pass, in the face of opposition that it would arm unscrupulous husbands to desert their wives and profit by the desertion. The Commission returned to the subject in its two hundred and seventeenth report, in 2009, and recommended the addition again, this time drawing on the Supreme Court’s own repeated observations. The Court had supplied its own pressure: in Naveen Kohli v Neelu Kohli, decided in 2006 and reported at (2006) 4 SCC 558, it granted a divorce on the ground of cruelty on facts of a marriage long dead, and recommended to the Union of India that it consider amending the Act to include irretrievable breakdown as a ground, observing that a marriage which exists only on paper serves no purpose.

The legislative attempt that followed is the one on which the whole argument turns. The Marriage Laws (Amendment) Bill of 2010, introduced in the Rajya Sabha, proposed to insert breakdown as a ground in both this Act and the Special Marriage Act, on proof that the parties had lived apart for a continuous period of not less than three years immediately preceding the petition, with safeguards that included a wife’s right to oppose the decree where it would result in grave financial hardship to her and the dismissal of the petition where the welfare of a child of the marriage so required. The Bill was passed by the Rajya Sabha in August 2013 after amendment. It was never taken up in the Lok Sabha, and it lapsed on the dissolution of the Fifteenth Lok Sabha in 2014. No successor bill has been enacted. The statutory position is therefore exactly what it was in 1955 on this point, and every general account that lists irretrievable breakdown among the grounds of divorce under this Act, as many do, is wrong.

The case for enacting the ground, stated at its strongest by its holders, is the case the Law Commission made twice and the Supreme Court made in Naveen Kohli: that a fault code applied to a faultless failure compels parties to fabricate or inflate allegations of cruelty and desertion, degrading the parties, the courts and the truth; that it prolongs proceedings across years in which neither party can remarry or settle; that it visits its worst costs on children, whose parents must perform hostility to earn an exit both parents want; and that the one-year separation already required for mutual consent, and the one-year non-resumption grounds of Section 13(1A), show the statute itself accepts separation over time as a legitimate measure of a marriage’s end. The case against, stated at its strongest by the women’s organisations that opposed the 1981 bill and pressed their objections through the 2010 Bill’s passage, and echoed judicially in the Supreme Court’s caution in K. Srinivas Rao v D. A. Deepa in 2013 that it is highly debatable whether, in the Indian situation, such a ground would be expedient at all, is that a unilateral no-fault exit in a society of profound economic inequality between spouses is a power given to the party who can afford to leave, ordinarily the husband; that maintenance law in India is weakly enforced and matrimonial property is not shared by any general rule, so that the wife expelled from a broken marriage by a breakdown decree faces destitution that the fault system, for all its cruelty, at least conditions on proof and on the court’s discretion under Section 23; and that the safeguard of an adequate financial settlement must therefore precede the ground rather than follow it, with the 2010 Bill’s hardship defence attacked by its critics as too narrow, since it placed the burden of opposing on the wife and measured her protection by grave hardship rather than by equal division. The verdict this article reaches is that both cases are honestly held and that the second has the better of the sequencing argument: the breakdown ground is just in principle and unsafe in isolation, and the thirty years of failure to enact it are explained less by legislative oversight than by the absence of the maintenance and property settlement that would make it safe. That is a reason the reform has stalled, recorded as such; it is not a reason to misdescribe the law, and the law is that the ground does not exist.

For three decades before the Constitution Bench settled the question, the Supreme Court’s own practice on breakdown was fractured, and the fracture is part of the record a reader should hold. One line of decisions dissolved marriages under Article 142 where the union was plainly dead, reasoning that the power to do complete justice answered the case before the Court whatever the statute omitted. Another line refused. In Vishnu Dutt Sharma v Manju Sharma, decided in 2009, the Court held that irretrievable breakdown is not a ground of divorce under the Act, that the earlier rulings dissolving marriages on that footing had not laid down the correct position, and that to grant the relief would be to add a ground the legislature had chosen not to enact, which is amendment and not adjudication. In Manish Goel v Rohini Goel, in 2010, the Court held that it could not grant a divorce in disregard of the substantive and procedural provisions of the Act, and could not dispense with the waiting period of Section 13B by constitutional fiat. Other benches continued to dissolve dead marriages in the exercise of the same power, among them the bench in R. Srinivas v R. Shametha in 2019, so that for a generation the outcome of a breakdown case in the Supreme Court depended on which line of authority the bench hearing it followed. The High Courts, meanwhile, were uniform in refusal: not holding the Article 142 power at all, they could do nothing for the dead marriage except strain the fault grounds, which is the precise mechanism by which breakdown was converted into cruelty in the courts below. It was to resolve that conflict, and to state the discipline governing the power, that the matter was referred to a bench of five.

What exists instead is the constitutional route, and its shape was settled by a Constitution Bench in Shilpa Sailesh v Varun Sreenivasan, judgment delivered on 1 May 2023, Sanjiv Khanna J. writing for a unanimous bench of five judges comprising Sanjay Kishan Kaul, Sanjiv Khanna, Abhay S. Oka, Vikram Nath and J. K. Maheshwari JJ. The references before the bench asked, among other questions, whether the Supreme Court may, in exercise of its power under Article 142(1) of the Constitution to pass such decree or make such order as is necessary for doing complete justice in any cause or matter pending before it, dissolve a marriage on the ground of irretrievable breakdown, and whether it may dispense with the procedure and the waiting period of Section 13B to do so. The Court held that it may. It held that the power under Article 142 is plenary within its sphere, that in a case of complete and irretrievable breakdown the Court may dissolve the marriage directly, may dispense with the statutory waiting period and the reference of the parties to the family court, and may do so even where one party opposes the dissolution, because to refuse on the ground of withheld consent would be to perpetuate a dead marriage by constitutional compulsion. It held equally that the power is discretionary and sparing: no party has a right to invoke Article 142 as a forum of first instance, the factors it enumerated include the duration of the marriage and of cohabitation, the period and circumstances of separation, with a separation of six years or more carrying particular weight, the nature of the allegations and the orders made in the proceedings between the parties, the attempts at settlement, and the parties’ ages, status and the economic and social consequences of dissolution, and the Court must be satisfied that the marriage is broken beyond repair and that amicable settlement is impossible. The decisions of two-judge benches that had held, as in Manish Goel v Rohini Goel in 2010, that the Court could not dissolve a marriage in contravention of the substantive provisions of the Act, stand displaced to the extent of the Constitution Bench’s ruling. No other court holds this power: the High Courts and the family courts remain confined to the statutory grounds, which is the precise sense in which the breakdown remedy in India is real and radically unequal in access, available to the litigant whose case reaches the apex court in a form that engages Article 142, and to no one else.

Maintenance, custody and the money consequences of the decrees

The Act’s financial provisions are three, and they operate at three different moments. Section 24 provides that where in any proceeding under the Act it appears to the court that either the wife or the husband has no independent income sufficient for her or his support and the necessary expenses of the proceeding, the court may, on application, order the respondent to pay the expenses of the proceeding and a monthly sum during the proceeding, having regard to the petitioner’s income and the respondent’s. The provision is formally sex-neutral, a point of real significance in the practice of the family courts, where husbands without income do claim and receive interim maintenance from earning wives. Proviso inserted by the Marriage Laws (Amendment) Act of 2001, Act 49 of 2001, directs that an application for expenses and interim maintenance shall, as far as possible, be disposed of within sixty days from service of notice on the respondent. Section 25 empowers the court, at the time of passing any decree or at any time subsequent to it, on application by either the wife or the husband, to order that the respondent pay the applicant, for her or his maintenance and support, a gross sum or a monthly or periodical sum for a term not exceeding the life of the applicant, having regard to the respondent’s income and property, the applicant’s income and property, the conduct of the parties and other circumstances, and the sum may be secured by a charge on the respondent’s immovable property. A Section 25 order may be varied, modified or rescinded where a change in the circumstances of either party is shown, and it is rescinded on the application of the party in whose favour it was made if that party has remarried, or, in the case of a wife, if she has not remained chaste, or, in the case of a husband, if he has had sexual intercourse with any woman outside wedlock. Section 26 empowers the court in any proceeding under the Act to pass interim orders, and to make provision in the decree, with respect to the custody, maintenance and education of minor children, consistently with their wishes wherever possible, and to modify or revoke such orders on application; the sixty-day disposal proviso of 2001 applies to custody applications as well.

These provisions do not stand alone in the life of a separated spouse, and the Act’s pillar must mark the boundaries rather than absorb the neighbours. Maintenance is also claimable under the general maintenance provision of the criminal procedure code and, for a Hindu wife, under the Hindu Adoptions and Maintenance Act of 1956, whose Section 18 grounds include the husband’s desertion and cruelty; the jurisdictions are concurrent and the choice among them is a matter of forum, quantum and enforcement that belongs to the maintenance articles of this series. Where the separation involves violence or the threat of it, the Protection of Women from Domestic Violence Act of 2005 supplies a parallel civil scheme of protection orders, residence orders and monetary relief that operates independently of any matrimonial petition, and the pillar article on the domestic violence statute, which deals with residence and maintenance in that scheme, is the correct next step for that reader. What the Hindu Marriage Act itself guarantees is narrower and worth restating exactly: a spouse without sufficient independent income can be kept in the litigation and supported during it under Section 24; a decree can carry a permanent settlement under Section 25 that survives the marriage and ends, in the events stated, with remarriage; and the children of the marriage can be provided for as to custody, maintenance and education under Section 26 in the same proceeding that decides the marriage. One further provision belongs with the money consequences because it is invoked in nearly every contested decree. Section 27 provides that in any proceeding under the Act, the court may make such provisions in the decree as it deems just and proper with respect to any property presented, at or about the time of marriage, which may belong jointly to both the husband and the wife. The section is narrow by design: it reaches property presented at or about the marriage which belongs jointly to both spouses, and it is not a general jurisdiction to divide matrimonial assets, of which Indian law has none within this Act. Property that is exclusively the wife’s, including her stridhan, is outside the section and must be recovered by the remedies the general law and the domestic violence statute supply, a boundary that litigants chronically misunderstand and that explains why the division of household assets on divorce is litigated under several statutes at once rather than under the marriage Act that dissolves the marriage. The absence of any matrimonial property regime is the silent premise of the breakdown debate: a legal system that cannot divide the assets of a marriage must think harder about the terms of exit than a system that can, and the Act’s silence on property is the strongest single explanation for the caution with which its exit law has been reformed.

The weakness of the scheme is not its text but its enforcement and its quantum, on which the record of the family courts is candid in the literature and which is the factual foundation of the objection to the breakdown ground recorded above.

The Hindu Marriage Act 1955 among the personal laws: boundaries and the civil alternative

The Act is one instrument in a system of personal laws, and its boundaries are propositions of law a reader needs in order to know whether it applies at all. It governs a marriage between two persons within Section 2. A marriage between a Hindu and a person who is not a Hindu, Buddhist, Jain or Sikh is not solemnised under this Act, and where the parties marry by Hindu rites in that situation the validity of the union has generated litigation the Act does not resolve in their favour. The alternative the legal system supplies is the Special Marriage Act of 1954, the civil statute under which any two persons may marry before a Marriage Officer regardless of religion, whose own conditions, notice regime and divorce provisions are the subject of its pillar article. A marriage solemnised under the Hindu Marriage Act is not converted into a civil marriage by registration, and a marriage solemnised under the Special Marriage Act is not governed by this Act’s grounds; the two statutes are parallel doors into marriage with different laws of exit behind each, and the comparison of the personal and secular routes, and of the succession consequences that follow the choice between them, is owned by the comparison article in this cluster, to which the reader deciding between the routes is directed rather than re-served here.

The boundary case that recurs in the courts is the marriage solemnised by Hindu rites between a Hindu and a person outside Section 2, most often a Christian. In Gullipilli Sowria Raj v Bandaru Pavani, decided in 2009 and reported at (2009) 1 SCC 714, the Supreme Court held that a marriage performed according to Hindu rites between a Hindu and a Christian does not become a marriage governed by this Act, since the Act applies only where both parties are within its application clause, and the union could not be treated as a Hindu marriage for the purposes of a prosecution built upon it. The proposition is severe and it is settled: the rites do not supply what the parties’ status withholds, and a couple across the boundary who marry by religious ceremony alone, without the civil form, may find that they have married under no statute at all, with consequences for status, maintenance and the legitimacy questions that Section 16 answers only within the Act. The remedy is the civil marriage, and the practical advice embedded in the case law is that the boundary is checked before the ceremony, not after it.

Two features of the boundary deserve emphasis because the general accounts blur them. The first is the saving of custom in Section 29(2), which provides that nothing contained in the Act shall be deemed to affect any right recognised by custom or conferred by any special enactment to obtain the dissolution of a Hindu marriage, whether solemnised before or after the commencement of the Act. A community with a proved customary mode of divorce, the panchayat dissolution and the deed of dissolution of the communities that practise them, may therefore end a marriage without any decree under Section 13, and the courts have recognised such customary divorces where the custom is pleaded and proved to the standard Section 3 requires. The saving is the Act’s acknowledgment that for a large part of the population it governs, divorce did not begin in 1955; what began in 1955 was a uniform court-based law of divorce standing beside the custom it declined to abolish. The second is the treatment of marriages solemnised before the Act. The Act’s remedial provisions, divorce, judicial separation and the rest, apply to marriages solemnised before its commencement as they apply to marriages solemnised after it, which is why a marriage of 1940 could be dissolved under Section 13 in 1960, and why the wife’s ground in Section 13(2)(i) was needed for the polygamous marriages the earlier law had permitted and the Act could not retrospectively annul. Entry into marriage before 1955 is judged by the law of its time; exit from it is judged by the Act.

Assessment: what the Act does, and the verdict

The Act’s achievements are structural and they are not small. It made the Hindu marriage monogamous as a matter of law and bigamy a matter for the criminal courts, ending a licence the classical law had given to one sex. It gave wives and husbands the same grounds and the same courts, ending the position in which a wife’s only remedies were the ones custom or charity supplied. It built a complete remedial ladder, from restitution through judicial separation to divorce, with maintenance and custody attached at every rung, and it protected the children of failed and void marriages by a legitimacy provision that removed from them the consequences of their parents’ status. It has been maintained with reasonable care: the capacity language modernised, the epilepsy reference removed, the leprosy ground repealed, the wife’s forum corrected, interim applications put under a disposal direction. As a code of entry and of fault-based exit, it is coherent, and the family courts administer it at a scale its drafters did not foresee.

The Act’s theory of exit can be stated in one sentence, and it organises everything in this article. Parliament enacted three theories of divorce without ranking them: the fault theory of 1955, under which a marriage ends because a spouse committed a listed wrong or suffers a listed disability; the consent theory of 1976, under which a marriage ends because both spouses ask the court to end it after a year apart; and the embryo of a breakdown theory, present since 1964 in the non-resumption grounds, under which a marriage ends because time has passed since an earlier decree without the parties coming together. What Parliament never enacted is the breakdown theory in its general form, divorce because the marriage has ended, proved by a period of separation alone. Every controversy in the modern law of this Act is a controversy at the joints of those three theories: the waiver cases ask how far the consent theory’s own timetable bends; the cruelty cases ask how much of breakdown the fault theory can be made to carry; the Article 142 cases ask whether a constitutional power may supply the fourth theory the statute lacks. A reader who holds the three theories apart will not be misled by accounts that merge them, and will see why reform proposals fail when they are argued as though only one theory were in the field.

The Act in the courts shows a pattern that an honest assessment must record in terms of direction rather than invented figures. Contested petitions take years, interim applications multiply within them, and the reconciliation the statute commands is attempted in proceedings whose structure rewards the performance of irreconcilability, since the party who concedes fault concedes the ground. The family courts established under the Act of 1984 have changed the forum’s character, with counsellors attached and an emphasis on settlement that the parent Act’s text does not describe, and a substantial share of the petitions filed end in settlement, withdrawal or conversion into mutual consent rather than in a contested decree, so that the fault code’s real function in the system is often to supply the pressure under which consent is negotiated. That is neither the design of 1955 nor an accident of administration; it is the system the two texts, the fault code and the consent provision, produce when they operate together on marriages that have failed in fact. The Act’s defenders are entitled to the point that a system which drives parties to negotiated exits is not obviously worse than a system of unilateral exit without a property settlement; its critics are entitled to the answer that a negotiation conducted under the threat of a decade of litigation is not the free consent the statute imagines, and that the weaker party’s settlement is discounted by exactly the delay the stronger party can afford.

Its failures are equally structural, and they cluster at the same point. The Act contains no answer to the marriage that has failed without fault, other than the adversarial alchemy by which breakdown is converted into cruelty through years of hostile litigation; it retains in restitution a remedy whose reconciling theory its practical uses contradict and whose constitutionality is under renewed challenge; its maintenance provisions promise support that enforcement routinely fails to deliver at the level the promise implies; and its text has been carried forward by judicial supplementation at the two points of greatest need, the waiver of the mutual consent waiting period by interpretation and the breakdown dissolution by constitutional power, neither of which is available on equal terms to every litigant the Act governs. The namable claim with which this article began can therefore be stated in its final form. The Hindu Marriage Act 1955 gives every Hindu in India a law of divorce, and it withholds from all of them the one ground that describes the commonest reason marriages end; the ground exists in the reports of the Law Commission, in the text of a lapsed bill, and in the plenary power of one court in the country, and it does not exist in the statute. A reform that enacts it, on the back of a maintenance and settlement regime strong enough to protect the spouse who did not choose the exit, remains the unfinished business of the Hindu Code settlement of 1955.

Frequently Asked Questions

Q: What conditions must a Hindu marriage satisfy?

Section 5 of the Hindu Marriage Act 1955 states five conditions. Neither party may have a spouse living at the time of the marriage. Neither party may be incapable of giving valid consent in consequence of unsoundness of mind, or be suffering from mental disorder of such a kind or to such an extent as to be unfit for marriage and the procreation of children, or be subject to recurrent attacks of insanity. The bridegroom must have completed twenty-one years and the bride eighteen. The parties must not be within the degrees of prohibited relationship, and must not be sapindas of each other, unless the custom or usage governing each of them permits the marriage.

Q: Is registration of a Hindu marriage compulsory?

Under the Hindu Marriage Act itself, no: Section 8 empowers state governments to make rules for entering the particulars of a marriage in a Hindu Marriage Register for the facility of proof, and it states that the validity of a marriage is in no way affected by omission to register it. The Supreme Court nevertheless directed in Seema v Ashwani Kumar in 2006 that marriages of all citizens should be made compulsorily registrable in their respective states, and states have framed rules under that direction. Registration is therefore an administrative obligation in most states and the principal documentary proof of a marriage, without being a condition of the marriage’s validity.

Q: What are the grounds for divorce under the Hindu Marriage Act?

Section 13(1) allows either spouse to petition on the grounds that the other party has had voluntary sexual intercourse with any person other than the spouse, has treated the petitioner with cruelty, has deserted the petitioner for a continuous period of not less than two years, has ceased to be a Hindu by conversion, has been incurably of unsound mind or suffers mental disorder of the specified kind and extent, suffers venereal disease in a communicable form, has renounced the world by entering a religious order, or has not been heard of as alive for seven years. Section 13(1A) adds non-resumption of cohabitation for a year after a decree of judicial separation or of restitution, and Section 13(2) gives a wife four further grounds of her own.

Q: How does divorce by mutual consent work?

Section 13B, inserted by the amendment of 1976, allows both parties to present a joint petition to the district court on the ground that they have been living separately for a period of one year or more, that they have not been able to live together, and that they have mutually agreed that the marriage should be dissolved. The parties then move the court together again, not earlier than six months and not later than eighteen months after the petition, and if the petition has not been withdrawn and the court is satisfied that a marriage was solemnised and the averments are true, it passes a decree of divorce. Consent must subsist until the decree; either party may withdraw it before the second motion.

Q: Can the six month waiting period for mutual consent divorce be waived?

Yes. In Amardeep Singh v Harveen Kaur, decided in 2017 and reported at (2017) 8 SCC 746, the Supreme Court held that the period in Section 13B(2) is directory and not mandatory, and that the court hearing the petition may waive it. The conditions are that the six months, taken with the one year of separation required by Section 13B(1), are already over before the first motion; that all efforts at mediation and conciliation have failed with no likelihood of success; that the parties have genuinely settled their differences, including alimony, custody and every other pending issue; and that waiting will only prolong their agony. The one year of separation itself cannot be waived.

Q: Is irretrievable breakdown a ground for divorce in India?

Not under the statute. Parliament has never enacted irretrievable breakdown as a ground in the Hindu Marriage Act, although the Law Commission recommended it in its seventy-first report of 1978 and its two hundred and seventeenth report of 2009, and the Marriage Laws (Amendment) Bill of 2010, which would have added it after three years of living apart, was passed by the Rajya Sabha in 2013 and lapsed without being taken up in the Lok Sabha. The Supreme Court may dissolve a marriage on this ground in exercise of its power under Article 142 of the Constitution to do complete justice, as a Constitution Bench held in Shilpa Sailesh v Varun Sreenivasan in 2023, but no other court may, and no party may claim that power as a matter of right.

Q: What does a decree of restitution of conjugal rights do?

A decree under Section 9 declares that a spouse who has withdrawn from the society of the other without reasonable excuse is to return to that society. It is not enforced by physical compulsion; execution proceeds against the judgment-debtor’s property under the Code of Civil Procedure. Its practical effects are that it records the court’s finding that the separation is without reasonable excuse, which bears on the withdrawing spouse’s maintenance claims, and that it starts a period after which either party may obtain a divorce: non-restitution of conjugal rights for one year or more after the decree is a ground of divorce under Section 13(1A), available even to the party whose failure to comply created the ground.

Q: What happens if a Hindu marries again after conversion?

The first marriage continues. Conversion to another religion does not dissolve a marriage solemnised under the Hindu Marriage Act; it gives the other spouse a ground of divorce under Section 13(1)(ii), but until a decree is passed the marriage subsists. A second marriage contracted during that subsistence is void, and the party who contracts it is guilty of bigamy under Section 494 of the Indian Penal Code. The Supreme Court settled the position in Sarla Mudgal v Union of India in 1995, holding that the second marriage of a Hindu husband after conversion to Islam, contracted without dissolving the first marriage, is invalid, and it reaffirmed that position against a challenge to it in Lily Thomas v Union of India in 2000.

Q: Who does the Hindu Marriage Act apply to?

Section 2 applies the Act to any person who is a Hindu by religion in any of its forms or developments, including a Virashaiva, a Lingayat and a follower of the Brahmo, Prarthana or Arya Samaj; to any person who is a Buddhist, a Jain or a Sikh by religion; and to any other person domiciled in India who is not a Muslim, a Christian, a Parsi or a Jew by religion, unless it is proved that the person would not have been governed by Hindu law in these matters. Converts to these religions are included, and the Act reaches Hindus domiciled in India who are outside its territory. It does not apply to members of a Scheduled Tribe unless the Central Government directs otherwise by notification.

Q: What makes a Hindu marriage void?

Section 11 declares a marriage null and void where it contravenes one of three of the conditions in Section 5: the monogamy condition, where either party has a spouse living at the time of the marriage; the prohibited relationship condition, where the parties are within the degrees of prohibited relationship and no custom governing each of them permits the marriage; and the sapinda condition, where the parties are sapindas of each other and no such custom permits it. A void marriage is treated as no marriage from the beginning, and either party may obtain a decree of nullity declaring it so. A bigamous marriage that is void under the section also attracts the punishment for bigamy through Section 17.

Q: On what grounds can a Hindu marriage be annulled as voidable?

Section 12 lists four grounds on which a marriage, valid until annulled, may be avoided by decree. The marriage has not been consummated owing to the impotence of the respondent. The marriage contravenes the mental capacity condition in Section 5(ii). The petitioner’s consent, or a guardian’s consent where that was required, was obtained by force or by fraud as to the nature of the ceremony or as to a material fact or circumstance concerning the respondent. The respondent was pregnant at the time of the marriage by some person other than the petitioner. Time limits and conduct bars in Section 12(2) apply, including a one-year limit from the cessation of force or the discovery of fraud, and a bar where the petitioner continued to live with the respondent with full consent after discovery.

Q: What are the sapinda and prohibited relationship rules, and when does custom override them?

Section 3(g) lists the degrees of prohibited relationship, which include lineal ascendants and descendants and their spouses, specified widows and widowers of close relations, and relations such as brother and sister, uncle and niece, aunt and nephew and the children of two brothers, two sisters, or a brother and sister. Section 3(f) defines sapinda relationship by ascent through common ancestors, extending to the third generation in the line of ascent through the mother and the fifth through the father. Section 5 forbids marriage within either category unless the custom or usage governing each of the parties permits a marriage between the two, so the custom must govern both parties, must be proved as certain, long observed and reasonable, and where it applies it is a complete answer.

Q: Which ceremonies make a Hindu marriage legally complete?

Section 7 provides that a Hindu marriage may be solemnised in accordance with the customary rites and ceremonies of either party. Where those rites include the saptapadi, the taking of seven steps by the bridegroom and the bride jointly before the sacred fire, the marriage becomes complete and binding when the seventh step is taken. The Act prescribes no ceremony of its own, so a community whose rites do not include the saptapadi marries validly by its own rites. In a dispute, the party asserting the marriage must prove the ceremonies by which it was solemnised, and proof of the saptapadi is required where it forms part of the applicable rites, a matter on which bigamy prosecutions have failed.

Q: What is judicial separation and how does it differ from divorce?

Judicial separation under Section 10 is a decree that suspends the obligation to cohabit without dissolving the marriage. Either party may petition on any of the grounds in Section 13(1), and a wife on her additional grounds in Section 13(2), since the amendment of 1976 aligned the grounds of the two remedies. The parties remain husband and wife: neither may remarry, obligations not dependent on cohabitation continue, and maintenance and custody orders may be made. The court may rescind the decree on the application of either party if it considers that just and reasonable. The decree also starts a clock, because non-resumption of cohabitation for one year or more after it is a ground of divorce under Section 13(1A) for either party.

Q: What extra grounds of divorce does a wife alone have under the Act?

Section 13(2) gives a wife four grounds not available to a husband. In the case of a marriage solemnised before the Act commenced, that the husband had married again before that commencement or that another wife of his from before that commencement was alive when her marriage was solemnised. That the husband has, since the solemnisation of the marriage, been guilty of rape, sodomy or bestiality. That a decree or order awarding her maintenance has been passed in her favour under the Hindu Adoptions and Maintenance Act or in a proceeding under the general maintenance provision of the criminal procedure code, and cohabitation has not been resumed for one year or more although she was living apart. That her marriage was solemnised before she attained fifteen and she repudiated it after attaining fifteen and before attaining eighteen.

Q: Can a divorce petition be filed within the first year of marriage?

Ordinarily no. Section 14 provides that no petition for divorce shall be entertained by any court unless one year has elapsed since the date of the marriage. The court may allow a petition to be presented before that year has run on the ground of exceptional hardship suffered by the petitioner or exceptional depravity on the part of the respondent, taking into account the interests of any children and the reasonable probability of reconciliation. If a decree is obtained on such an application by misrepresentation or concealment of the nature of the case, the court may make the decree conditional on its not taking effect until the year has expired, or may dismiss a later petition founded on the same facts without prejudice.

Q: When can divorced persons marry again?

Section 15 provides that when a marriage has been dissolved by a decree of divorce and either there is no right of appeal against the decree, or, if there is such a right, the time for appealing has expired without an appeal having been presented, or an appeal has been presented but has been dismissed, it shall be lawful for either party to the marriage to marry again. A remarriage contracted while an appeal is pending or still available is contracted at risk, since the decree dissolving the first marriage may be reversed; the section ties the liberty to remarry to the finality of the decree, and parties are ordinarily advised to wait until the appeal period has run out or any appeal has been disposed of.

Q: What maintenance and expenses can a spouse claim while a case is pending?

Section 24 provides that in any proceeding under the Act, where it appears to the court that either the wife or the husband has no independent income sufficient for her or his support and the necessary expenses of the proceeding, the court may order the respondent to pay the expenses of the proceeding and a monthly sum during the proceeding, having regard to the incomes of both parties. The provision is available to a husband as well as to a wife. An amendment of 2001 directed that an application under the section should, as far as possible, be disposed of within sixty days of service of notice on the respondent. Maintenance for the period after the decree is a separate matter governed by Section 25.

Q: Are children of a void or annulled marriage legitimate?

Yes, within the limits Section 16 states. The section provides that notwithstanding that a marriage is null and void under Section 11, any child of the marriage who would have been legitimate if the marriage had been valid shall be legitimate, whether the child was born before or after the amendment of 1976 and whether or not a decree of nullity is granted or the marriage is held void on a petition under the Act. Children of a voidable marriage annulled by decree receive the same protection. Section 16(3) bounds the rule: the child acquires no rights in the property of any person other than the parents by virtue of the section, so the legitimacy carries inheritance from the parents and not against the wider family.

Q: Can a Hindu marriage be dissolved by custom without a court decree?

Yes, where a genuine custom supports it. Section 29(2) saves any right recognised by custom, or conferred by a special enactment, to obtain the dissolution of a Hindu marriage, whether the marriage was solemnised before or after the Act commenced. Communities with an established customary mode of divorce, such as dissolution by a caste panchayat or by a deed of dissolution where the community’s usage recognises it, may therefore end a marriage without a petition under Section 13. The custom must be pleaded and proved to the standard the Act’s definition requires: a rule continuously and uniformly observed for a long time, certain, not unreasonable and not opposed to public policy. A private agreement between spouses, without such a custom behind it, dissolves nothing.