For seven decades, Indian courts have been performing a function that no comparable court in the world undertakes. When a religious community seeks constitutional protection for a particular practice, courts first ask whether that practice is “essential and integral” to the religion. Only practices that pass this test receive shelter under Articles 25 and 26, which guarantee religious freedom—the former to individuals, the latter to religious denominations—subject in each case to public order, morality, and health.

This doctrine, known as the Essential Religious Practices test or ERP, requires secular judges to make theological determinations about the core content of religious faiths.

Solicitor General (SG) Tushar Mehta’s written submissions to the nine-judge Constitution Bench span three volumes and more than 180 pages. They mount the most sustained structural challenge to ERP since the doctrine was fashioned. In 1962, a judgment converted a descriptive word into a constitutional threshold. That conversion, the SG argues, is the source of seven decades of doctrinal error.

The reference arises from the Supreme Court’s judgment of September 28, 2018, when a five-judge bench ruled 4:1 that women of menstruating age could not be barred from the Sabarimala temple in Kerala. Review petitions followed. On November 14, 2019, a bench referred seven constitutional questions to a larger bench. A nine-judge Constitution Bench led by Chief Justice Surya Kant has been hearing those questions since April 7, 2026.

The wrong turn of 1962

The foundational case on religious freedom is Commissioner, Hindu Religious Endowments, Madras v Sri Lakshmindra Thirtha Swamiar of Sri Shirur Mutt (1954). That bench used the word “essential” descriptively: to identify the practices a religion itself regards as central, and to protect those practices from state interference. The word served as a shield for religious autonomy.

Justice Gajendragadkar’s bench in Durgah Committee, Ajmer v Syed Hussain Ali (1962) made two innovations that Shirur Mutt had not sanctioned. First, it converted “essential” from description into a threshold: any practice must now be “essential and integral” to the religion before receiving constitutional protection. Second, it excluded all practices characterised as “superstitious” from Article 25’s ambit, on the ground that protecting them would “defeat the very protection which Part III intends to afford.” These two steps created ERP in the form Indian courts now apply.

The submissions argue that neither innovation was authorised by Shirur Mutt, which simply protected what a religion regards as its own. The Durgah Committee converted that protection into a certification process, with courts acting as theological arbiters. The submission’s answer to the obvious objection is sincerity. Courts can ask whether a community sincerely holds a practice as religiously required. Assessing whether the practice is theologically central lies beyond judicial competence. Comparatively, the point is confirmed by Church of the New Faith (1983), in which the Australian High Court—whose earlier decision Shirur Mutt cited as authority for ERP—held that the state has “no prophetic role in relation to religious belief.” The country whose case fathered the doctrine moved in the opposite direction.

Part II raises an argument that could terminate the challenge before any question of religious doctrine is reached. Article 13 of the Constitution empowers courts to strike down any “law” that abridges fundamental rights. Its definition of law includes customs and usages, but only those “having the force of law.” A custom acquires that force when its observance is mandated through a public legal norm, or when its violation attracts legal rather than purely spiritual consequences. A voluntary denominational observance, practised as a matter of faith, falls entirely outside Article 13’s scope.

The submissions distinguish three tiers of religious practice: purely voluntary denominational observance; practice backed by codified personal law; and practice translated into a statutory instrument. Courts may scrutinise only the second and third tiers. The 2018 majority never asked which tier Sabarimala’s practice occupied. That omission, the SG argues, vitiated everything that followed.

Rule 3(b) of the Kerala Hindu Places of Public Worship Rules, 1965, was the law the 2018 bench struck down. The submissions meet this counter at a different level: even granting that Rule 3(b) was the reviewable instrument, the court’s jurisdiction extended to that rule alone, not to the underlying voluntary practice.

The deity’s legal personality

Part II advances a second structural argument from settled Indian property law. Under Indian law, a Hindu deity is a juristic entity: a legal person capable of owning property and suing in court. The law treats a deity as a perpetual minor. The Privy Council established this in Pramatha Nath Mullick v Pradyumna Kumar Mullick (1925). The Supreme Court confirmed it in Bishwanath v Shri Thakur Radha Ballabhji (1967). A court cannot evaluate the personal attributes of a perpetual minor on grounds of irrationality. By the same logic, it cannot evaluate a deity’s attributes on any criterion of rationality.

The Sabarimala temple is dedicated to Lord Ayyappa as a Naisthika Brahmachari, meaning one who has taken a permanent vow of celibacy. Elsewhere, Lord Ayyappa is worshipped in forms that admit women freely. The restriction flows from the specific consecrated character of the deity at this shrine. The denomination’s understanding of that character is the only understanding the law is competent to accept. The 2018 majority examined whether the celibate character was “essential” to the religion. The submissions call this a category error: the deity’s legal personhood makes the worshipping community’s understanding of its own deity conclusive.

Opposing counsel will argue that this principle belongs to property law and cannot govern fundamental rights proceedings. The submissions answer that juristic personhood is a structural principle of Indian law without domain limits. If the denomination’s understanding of a deity’s attributes is authoritative in property matters, the same self-understanding carries equal weight in constitutional proceedings.

Article 25(2)(b) of the Constitution empowers the state to open Hindu religious institutions of public character to all classes and sections of Hindus. The 2018 majority used this as one basis for overriding the Sabarimala restriction. Part II challenges that reading, drawing on the Constituent Assembly debates. The founding fathers were conscious that certain communities were denied temple entry on grounds of birth, that is, caste. Article 25(2)(b) was their constitutional response to that specific wrong. Wherever the framers intended sex or gender as a protected ground, they named it explicitly: Articles 15, 16, and 325 name sex. The temple-entry provision does not.

The opposing reading holds that women constitute a “section of Hindus” to whom the provision equally applies. The submissions meet this with the Constituent Assembly record: K.M. Munshi’s original draft, and the debates consistently linked the provision to caste discrimination. Gender was already addressed through Articles 14, 15, and 16. Article 25(2)(b) was crafted for a different, and more specific, wrong.

A related gap concerns Article 26, which gives every religious denomination the right to manage its own affairs. Courts have applied a restrictive definition of “denomination”, while the companion phrase “or a section thereof” has received no independent analysis. The SG argues that “section thereof” was inserted to protect smaller communities and sampradayas (religious traditions) with sincerely held beliefs but no formal hierarchy, and that the Sabarimala Ayyappa tradition may qualify for independent protection under Article 26(b).

A doctrine with no counterpart

Part III closes the structural argument with a comparative survey. In Employment Division v Smith (1990), the United States Supreme Court held that judicial inquiry into the centrality of religious practices is “akin to the unacceptable business of evaluating the relative merits of differing religious claims.” In Thomas v Review Board (1981), the same court held that intra-faith disagreements are “singularly ill-suited” to judicial resolution: sincerity of belief, not theological centrality, is the proper inquiry. In Leyla Sahin v Turkey, the European Court upheld a headscarf restriction on proportionality grounds, without asking whether the headscarf was essential to Islam. ERP has no counterpart in any of these jurisdictions.

ERP is most intrusive when applied to a religion with no founder, no single canonical scripture, and no ecclesiastical authority. Hinduism is precisely such a religion. The doctrine imposes its heaviest burden on the faith whose internal plurality makes compliance most difficult.

The written submissions have structured their argument on several levels simultaneously. At the jurisdictional level, they ask whether a voluntary religious observance qualifies as a “law” under Article 13. At the doctrinal level, they ask whether a 1962 judgment corrupted the foundational 1954 case, and whether ERP survives comparison with any comparable constitutional order. At the hermeneutic level, they ask whether the attributes of a juristic deity are for the court to evaluate. Each level, if resolved in favour of the Union of India, renders the next unnecessary.

Also Read | Sabarimala’s season of controversies

Also Read | For the love of god: Why Kerala’s Global Ayyappa Sangamam is drawing BJP’s ire