Due process, smuggled in: Article 21 after Maneka

The Indian Constitution, as drafted, contains no due process clause. The framers considered the American formulation, weighed it, and removed it. Article 21 in the document of 1950 reads: 'No person shall be deprived of his life or personal liberty except according to procedure established by law.'¹ The phrase 'due process of law' was, in the words of B Shiva Rao's documentary record, deliberately deleted. Three decades later, in Maneka Gandhi v Union of India, a seven-judge bench held that the procedure contemplated by Article 21 had to be 'fair, just and reasonable'. Within five years the same Court had extended that test to the substance of penal statutes. The clause that the framers had narrowed at the drafting table now carries both procedural and substantive due process by judicial reading. The transition is constitutional history. The question this essay surveys is what was lost in the route the doctrine took.

What the framers refused, and why

The story has two protagonists who never met in court. The first is Sir Benegal Narsing Rau, constitutional advisor to the Constituent Assembly, who in September 1946 and again in April 1947 expressed doubts in his notes about whether the American due process clause should be introduced into India's proposed constitution. His fear was institutional rather than ideological. The clause, he wrote, had given American courts excessive discretion over legislation, and might empower Indian courts to obstruct what he called 'beneficient social legislation' — chiefly the planned land redistribution that was politically central to the Congress programme.

The second is Justice Felix Frankfurter, whom Rau met in the United States sometime between October and December 1947. Frankfurter advised Rau to delete the phrase 'due process of law' on two grounds: that it was 'undemocratic' because it gave a few judges the power of vetoing legislation enacted by the representatives of the nation, and that it imposed an 'unfair burden' on the judiciary. Frankfurter's advice almost certainly reinforced rather than created Rau's view. But it gave the deletion a foreign endorsement that the Drafting Committee found persuasive. By 13 December 1948, after a sustained debate in which numerous members moved amendments to reintroduce the phrase, the Constituent Assembly fixed the present text.² Article 21, in the version that came into force on 26 January 1950, contained the Japanese Constitution's wording at Article 31 rather than the American.

Three other changes accompanied the deletion. The right to property was dropped from the clause, on the same anxiety about social legislation. The word 'liberty' was qualified with 'personal' — Rau's specific intervention, born of his fear that an unmodified 'liberty' would be read so broadly as to capture even commercial regulation in the manner of the Lochner court. And as a concession to those who had argued for due process, a separate provision (what became Article 22) inserted express procedural safeguards against arbitrary arrest. The architecture of Article 21 was thus deliberately narrow. It was a guarantee against executive lawlessness, not against legislative unfairness.

It is important to record the texture of the floor debate, because the standard account flattens it. The deletion of due process was not consensus business. Numerous amendments were moved to reinstate the clause. Only Alladi Krishnaswami Ayyar spoke strongly in favour of the deletion, predicting that 'due process' would 'serve as a great handicap for all social legislation' and create 'judicial vagaries in the moulding of law'. Even Ambedkar was, by Granville Austin's reading, ambivalent. Kazi Syed Karimuddin captured the consequence the framers were knowingly accepting: under 'procedure established by law', he said, 'as soon as the procedure is complied with, there will be an end to everything and the judges will be only spectators'.¹¹ That was the bargain. Frankfurter, writing in the Harvard Law Review in 1955, credited Rau by name with the deletion.³

The Gopalan settlement

The first sustained occasion to interpret Article 21 came almost immediately, in the very first fundamental rights case to reach the Supreme Court. AK Gopalan v State of Madras challenged the Preventive Detention Act 1950 on grounds that included an argument that 'procedure established by law' should be read to import natural justice. A six-judge bench rejected that reading, five-to-one on the question of construction.⁴ Five of the six judges discussed the American due process clause. Each rejected its importation. Kania CJ pointed to the three deliberate departures from the American text. Sastri J wrote that to read 'procedure established by law' as procedural due process would be to revive the very 'subtle and elusive criteria' that the framers had taken pains to avoid. Das J went further: 'it is common knowledge that our Constitution-makers deliberately declined to adopt the uncertain and shifting American doctrine of due process of law'.

Only Fazl Ali J dissented on the construction. His reasoning was learned but ultimately speculative. The phrase 'procedure established by law' had reached the Indian draft via Japan; the Japanese constitution had been drafted under American influence; American courts had read 'law' in the due process clause to mean procedural due process; therefore 'procedure established by law' meant procedural due process. The chain has the elegance of a syllogism and the brittleness of one. The majority found it unpersuasive.

What Gopalan settled, alongside the construction, was a structural reading of Part III. The judgment held that Articles 19, 21 and 22 occupied separate, non-overlapping fields. A law which validly took away personal liberty under Article 21 could not be tested against the Article 19 freedoms; a person whose Article 19 rights had been suspended through a valid Article 21 deprivation could not complain. This silos-of-rights formalism would, in time, prove the more vulnerable holding. By the late 1960s it was already coming apart in the property-rights cases, and in RC Cooper v Union of India (1970) an eleven-judge bench overruled it. After Cooper, a single state action could be assessed against multiple fundamental rights provisions read together.

The point worth dwelling on is that Cooper did not require Maneka. The silos-of-rights problem could be — and was — fixed without reading due process into Article 21. The interpretive question that Maneka reopened in 1978 was a different one: not whether Article 21 could be read alongside Article 19, but whether the procedure under Article 21 had to satisfy substantive standards of fairness. On that, Gopalan had been clear, the framers had been clearer, and the textual position was settled.

What Maneka actually did

The facts of Maneka Gandhi v Union of India are familiar. Indira Gandhi's daughter-in-law had her passport impounded under section 10(3)(c) of the Passports Act 1967, in the 'public interest', without reasons being supplied. She challenged the order under Articles 14, 19 and 21. The seven-judge bench restored her passport and, in doing so, reformulated the meaning of Article 21.

Bhagwati J's leading judgment posed the question crisply: does Article 21 'merely require that there must be some semblance of procedure, howsoever arbitrary or fanciful, prescribed by law before a person can be deprived of his personal liberty, or that the procedure must satisfy certain requisites in the sense that it must be fair and reasonable?' The majority answered the second way. The procedure had to be 'fair, just and reasonable, not fanciful, oppressive or arbitrary'.⁵ Chandrachud J and Krishna Iyer J wrote substantially the same words in their concurring judgments. The phrasing — Abhinav Chandrachud has noted — is 'remarkably reminiscent' of the language of the American Supreme Court in the Lochner era.

What the Court did not do, and ought to be acknowledged, was overrule Gopalan by name on the silos point. That had already been done in Cooper. What the Court did do was hold, in plain terms, that procedural due process was now part of Article 21, and that the test of reasonableness flowed in from Article 14 because — Bhagwati J reasoned — Article 14 itself contained an inherent reasonableness requirement, and Articles 14, 19 and 21 had to be read together as a unified scheme of protection. The integration of the three articles is the real doctrinal innovation. The 'fair, just and reasonable' test is the mechanical consequence.

It is also worth saying — because the standard account underplays it — that the bench knew exactly what it was doing. Three of the seven judges who decided Maneka (Beg, Chandrachud and Bhagwati JJ) had been in the majority in the Habeas Corpus case two years earlier. The lesson the Court drew from that majority — Khanna J's dissent having been the only opinion to come out of Habeas Corpus with reputation intact — was, as Chandrachud has put it, that the Constitution did not always have to be interpreted with strict fidelity to its text. The Emergency had been the Court's failure. Reading expansively was its rehabilitation. Maneka was, in this sense, partly an act of constitutional repair.

From procedure to substance, in five years

The strict logic of Maneka applied only to procedural laws. The case was about the procedure for impounding a passport. The reasoning extended to procedure for arrest, procedure for trial, procedure for sentence. It did not, in terms, hold that the substantive content of a penal statute could be tested for fairness under Article 21. That extension came in stages.

The first step was Sunil Batra v Delhi Administration, decided seven months after Maneka. The challenge was to section 30(2) of the Prisons Act 1894, which provided that a prisoner 'under sentence of death' was to be confined 'in a cell apart from all other prisoners'. The question was whether this permitted solitary confinement. The Court read the provision down: a prisoner was 'under sentence of death' only after the entire appellate process had been concluded and clemency was denied; and the words of the section did not in themselves require solitary confinement. Sunil Batra was a procedural due process decision — the Prisons Act provision regulated treatment, not punishment, and the IPC sections that defined solitary confinement as a punishment were not before the Court.⁶ Its importance lay elsewhere. It was the moment at which Krishna Iyer J wrote, in concurrence: 'True, our Constitution has no due process clause but, in this branch of law, after Cooper and Maneka Gandhi, the consequence is the same.' That sentence is the obituary of the framers' choice, written by a sitting judge in 1978.

The substantive extension followed in 1980. Bachan Singh v State of Punjab challenged section 302 of the Indian Penal Code — a substantive provision, prescribing the death penalty for murder — on the ground that it violated Article 21. The majority could have refused jurisdiction on the basis that Maneka spoke only to procedure. They did not. They tested section 302 on the touchstone of Article 21, found it constitutional on the merits, and in doing so reformulated the article. Their majority paraphrase, at paragraph 136, reads Article 21 as: 'No person shall be deprived of his life or personal liberty except according to fair, just and reasonable procedure established by valid law.'⁷ The phrase 'valid law' is doing very specific work. It means: substantive content, not merely the procedure of enactment.

Bhagwati J's dissent in Bachan Singh, delivered two years after the majority, made the move explicit. 'The word procedure in Article 21', he wrote, 'is wide enough to cover the entire process by which deprivation is effected and that would include not only the adjectival but also the substantive part of the law.' Three years after Maneka, by the author of Maneka, the doctrine had grown a second leg. By 1983, in Mithu v State of Punjab, the Court was striking down section 303 IPC — the mandatory death penalty for life convicts — on what is, in substance, both a procedural and a substantive due process ground. 'A savage sentence is anathema to the civilized jurisprudence of Article 21,' the Chief Justice wrote.⁸ Hypothetical though the example was — the Court spoke of a law that punished theft by the cutting of hands — the proposition was unmistakable: substantive criminal statutes could now be reviewed on Article 21 for substantive fairness.

By the time the Court in Selvi v State of Karnataka (2010) used the phrase 'substantive due process' as a self-conscious doctrinal category — assessing whether narco-analysis and brain-mapping techniques violated Article 21 — the question of whether the doctrine existed in Indian law had stopped being asked.⁹ By K.S. Puttaswamy v Union of India (2017), recognising privacy as intrinsic to Article 21, the integrated reading of Articles 14, 19 and 21 was treated as elementary doctrine, not as innovation.¹⁰ The textual deletion of December 1948 was, by then, more or less a historical curiosity.

The pace of the move bears emphasis. Five years separate the four cases that completed the doctrinal transition — Maneka in January 1978, Sunil Batra in August 1978, Bachan Singh in May 1980, Mithu in 1983. Each step was small. Each was justified, in turn, by the preceding step. Sunil Batra drew its authority from Maneka; Bachan Singh drew its from the integrated reading the post-Maneka Court had begun to deploy; Mithu drew its from Bachan Singh's extension. None of the four judgments characterised itself as constitutional invention. Each presented the move as the natural working-out of the previous one. This is how doctrinal accretion works, and there is nothing improper about it. But the cumulative result was a constitutional change of the first order — the importation of a clause the framers had specifically rejected — accomplished without anyone, at any single point in the line, having to defend the change as a change.

The strongest defence of the route taken

The defence of Maneka's reading is not weak. The most serious version of it has two parts.

The first is that Gopalan's silos-of-rights formalism had collapsed by the late 1960s and that the Court needed a new architecture. RC Cooper had established that fundamental rights overlapped, but it had not specified a test for what the Article 21 protection actually required. Without a substantive content for 'procedure established by law', Article 21 reduced to a guarantee that the executive would not act without statutory cover — a guarantee of legality, not of fairness. In a constitutional system that had just survived an Emergency in which laws were enacted that authorised arbitrary detention without judicial scrutiny, that minimum standard plainly was inadequate. Maneka's reasonableness reading filled the space that Gopalan's collapse had opened. On this view, the reading was not invention but engineering necessity.

The second part of the defence runs deeper. The framers had, the argument goes, themselves been ambivalent. The deletion of due process was a contested vote, not a consensus. Even Ambedkar wavered. The framers were worried about a particular kind of due process — the Lochner-style invalidation of social welfare legislation — and that worry has not, in the event, materialised. The Indian Supreme Court has not used Article 21 to strike down economic regulation on liberty-of-contract grounds. It has used it to expand free legal aid for undertrials in Hussainara Khatoon,¹² to humanise prison conditions in Sunil Batra, and to constrain the death penalty. The fears that drove the deletion were specific; the doctrine that emerged after Maneka has not honoured those fears. It has done something the framers might, on reflection, have approved of.

This is a strong case. It deserves to be met on its own terms, not waved away. There is no good textual answer to the engineering point — Article 21 read narrowly is, after the Emergency, a thin protection. There is no good intentionalist answer to the ambivalence point — the framers' silence is not evidence of consent, but it is not evidence of opposition either.

What the route cost

The doubt the essay wants to leave open is not whether the doctrine is defensible. The doubt is about the route by which it arrived.

India has had several constitutional amendments to its fundamental rights chapter — the First, the Twenty-Fourth, the Twenty-Fifth, the Forty-Second, the Forty-Fourth. The Forty-Fourth Amendment (1978), passed in the same year as Maneka, made significant changes to Articles 19 and 22 and removed the right to property from the fundamental rights chapter altogether. The mechanism for revisiting the deliberate choices of December 1948 was, in 1978, an active and recently-used part of the constitutional system. Parliament could have revisited the framers' deletion of due process. It did not. The Court did the work instead.

The substitution of judicial reading for democratic deliberation has costs that compound over time. Three are worth naming.

The first is interpretive. A doctrine added by amendment carries its content on its face. A doctrine read in by the judiciary acquires its content case by case. The 'fair, just and reasonable' formula has no internal limit. Once the Court holds — as it did in Maneka — that any procedure failing the test is unconstitutional, the meaning of the test becomes whatever the Court of the day decides it means. The expansion of Article 21 to include the right to livelihood, the right to shelter, the right to a clean environment, the right to education and the right to health is a real achievement of Indian constitutional law. It is also entirely judge-made content layered on a single sentence the framers had drafted to do narrow work. The text and the doctrine have, in some sense, parted ways.

The second cost is institutional. The framers of 1948 did not delete due process because they were against fairness. They deleted it because they were against the institutional implication of due process — that an unelected court would have the last word on whether a statute was fair, on a standard the court itself supplied. That institutional implication is now the settled fact of Indian constitutional law. It was settled not by the Constituent Assembly, not by Parliament, not by referendum, but by a seven-judge bench in 1978 and a five-judge bench in 1980. There is something uncomfortable in the asymmetry. The democratic decision was reversed by a non-democratic body using powers that the democratic decision had not granted.

The third cost is more diffuse, but real. A constitution acquires its authority partly from the visible record of how it came to be. When the text says one thing and the operative doctrine says another, the constitution becomes a document that needs to be explained rather than read. The lay citizen reading Article 21 today encounters a sentence about 'procedure established by law' and is told that the operative meaning is 'fair, just and reasonable procedure established by valid law'. The two are not the same sentence. To follow the move from one to the other requires reading Maneka, Sunil Batra, Bachan Singh, Mithu, the integration of Articles 14 and 19, and a body of doctrine accumulated over forty-five years. The Constitution as text and the Constitution as doctrine have become two different documents.

There is a fourth cost worth recording, even though it cuts in the opposite direction from the others. The route taken did spare India a particular hazard. A formal due process clause introduced by amendment in 1978 — under the Janata government, in the same political moment as the Forty-Fourth Amendment — would have come with the Forty-Second Amendment's overhang, with the Emergency's residue, with all the politics of that decade attached. It might have been drafted narrowly, drafted defensively, or drafted to serve a particular government's reading of fairness. The judicial route insulated the doctrine from those moments. Maneka was not the work of a Parliament; it was the work of a Court that had just learned, painfully, what an unrestrained Parliament could do. There is a real argument that the route protected the substance. It is an argument that has to be weighed against the deliberative cost. We have not yet, as a constitutional culture, settled how to weigh it.

Where the question rests

None of this is an argument for going back. It would be both impossible and unwise to undo Maneka. The constitutional architecture of Indian rights protection has been built on its foundation for nearly half a century. Hussainara Khatoon, Sunil Batra, Olga Tellis, Vishaka, Puttaswamy — none of them survives a return to Gopalan. The doctrine is doing real work that the constitutional text, as drafted, would not have done.

The argument is narrower. It is that we should be honest about what happened, and about what we lost in the manner of its happening. The framers of 1948, in the considered exercise of constitutional drafting, refused due process. The Court of 1978, in the considered exercise of constitutional adjudication, restored it. Both moves were defensible on their own terms. The route mattered, and continues to matter. A constitutional culture that lets its courts rewrite contested choices — even contested choices the courts get right — gradually relocates the seat of constitutional authorship. After enough of those moves, the question of who wrote the Constitution becomes harder to answer.

The unresolved question, fifty years on from Maneka's aftermath, is whether the gain in protection has been worth the loss in deliberative legitimacy — and whether, when the next contested constitutional choice comes, we will let the framers' equivalents of B.N. Rau and Alladi Krishnaswami Ayyar settle it, or whether we will let it be settled by a bench of seven.

Get the week's biggest decisions, distilled.

A weekly digest for lawyers, students and the legally curious. Every Friday morning. Free.