Fifty Years of ADM Jabalpur: India's Most Honest Judgment

On 28 April 1976, four judges of the Supreme Court of India held that during the operation of the Emergency proclaimed under Article 352, and a Presidential Order issued under Article 359, no person had locus standi to move a High Court for habeas corpus on the ground that his detention was illegal, mala fide, ultra vires, or for any other reason at all.¹ The country has spent fifty years explaining why the judgment was wrong. The 44th Amendment said so, in 1978. The Supreme Court, in Puttaswamy, said so in 2017. Every law-school syllabus since has said so. The judgment has the rare status of being almost universally agreed to be the worst of its kind in independent India. I want to argue that the unanimity is the problem. ADM Jabalpur was the most honest judgment the Indian Supreme Court has ever delivered. The four judges in the majority read the constitutional text the executive had handed them, and they said what that text actually authorised. The dissent of Justice H.R. Khanna was, and remains, the only intellectually serious answer the law produced. Everything between — the 44th Amendment, Maneka, the formal repudiation in Puttaswamy — is, in part, constitutional consolation. It put distance between the country and a judgment it could not bear to keep without dismantling the architecture that made the judgment possible.

The settled view, taken at its strongest

The settled view is that Shivakant Shukla was a moral collapse, that the four judges who delivered the majority opinion were variously cowardly, ideologically captured, or institutionally intimidated, and that the legal materials available to them did not in fact require the conclusion they reached. On this view, the dissent of Justice Khanna shows that a different decision was possible on the same text, and the majority's failure to deliver that decision is what makes the judgment a disgrace. The post-Emergency corrections — the 44th Amendment's exclusion of Articles 20 and 21 from any future Presidential Order, the rewriting of Article 352, the doctrinal advance in Maneka, the explicit overruling in Puttaswamy — together close the book on Shivakant Shukla and on the constitutional logic that produced it.⁴

I want to state this view at its strongest, because the polemic I am about to mount against it is not a polemic against any of the people who hold it. It is a polemic against what holding it allows the country to avoid. The strong form of the settled view has three components. Each is true, and each is incomplete.

The first component is moral. The judgment authorised, in practical effect, the detention of tens of thousands of people without judicial review, including political opponents, journalists, trade-union leaders, and ordinary citizens against whom no charge was ever brought. Many died in custody. Some were tortured. The Supreme Court of a constitutional democracy gave the executive a clear field to do this, and did so by an opinion whose dominant register was textualist deference. That the Court did this matters and will keep mattering. The instinct to call the judgment a disgrace is correct. The judgment is a disgrace. I am not interested in defending the four judges as people. I am interested in something else.

The second component is institutional. The Court that decided Shivakant Shukla had been packed and superseded. Justices Shelat, Hegde, and Grover had been bypassed for the office of Chief Justice in 1973 after the Kesavananda decision in punishment for an inconvenient ruling. Justice Khanna would himself be superseded in January 1977, with Beg J elevated over him to the office of Chief Justice, and would resign within days.¹³ The Court was not an independent institution in the strong sense in which one would hope. The structural pressure on the bench is part of what produced the judgment. The settled view is correct that this pressure was operative and that it should not be forgotten.

The third component is doctrinal. Shivakant Shukla read Article 359 as suspending not merely the enforceability of Article 21 but also any underlying right to life and personal liberty that could be claimed independent of the Article. Khanna J's dissent argued that even on the assumption that Article 21 was procedurally suspended, the State did not thereby acquire a positive power to deprive citizens of life or liberty without legal authority. The Supreme Court in Puttaswamy, four decades later, accepted the dissent's reasoning and overruled the majority.⁵ The settled view is correct that the doctrinal repair has been done, and that the judgment is no longer good law on any reading.

Take these three components together and the conclusion writes itself. The judgment was a moral failure, produced under institutional pressure, since corrected doctrinally and constitutionally. The country should remember it as a warning and move on.

I disagree with the moving-on. I disagree, more specifically, with the implicit claim that the doctrinal and constitutional repairs since 1976 have done the structural work the judgment exposed as needing to be done. The repairs have placed the country at a distance from the judgment. They have not dismantled the architecture that produced it. That distinction is what this essay is about.

What the four judges actually said

It is now nearly impossible to read the majority opinions in Shivakant Shukla without a presumption of bad faith. I want to ask the reader, for the length of this section, to suspend that presumption. The four opinions are not a single chorus. Ray CJ's opinion is the most political, Beg J's the most textualist, Chandrachud J's the most institutional, and Bhagwati J's the most philosophically anguished. What they share is a reading of the relevant constitutional text that I will now state plainly.

Article 352 of the Constitution, as it then stood, permitted the President to proclaim an emergency on his satisfaction that the security of India was threatened by 'war, external aggression, or internal disturbance'. Article 358 provided that during the operation of such a proclamation, the freedoms guaranteed by Article 19 stood suspended automatically. Article 359 went further: it permitted the President, by order, to suspend the right of any person to move any court for the enforcement of any of the fundamental rights conferred by Part III, except as the order might specify.² The Presidential Order of 27 June 1975 was issued under that provision. It suspended the right to move any court for the enforcement of the rights conferred by Articles 14, 21 and 22 for the duration of the Emergency.

The High Courts before whom habeas corpus petitions were filed under the Maintenance of Internal Security Act took the view that they could nonetheless entertain petitions challenging the legality of detentions on grounds other than the suspended Articles — for instance, that the detention order was outside the four corners of the statute, or that it was mala fide, or that it lacked any factual basis. The 1962 Presidential Order, issued during the war with China, had been narrowly drawn, and the Court in Makhan Singh v State of Punjab had read it down to leave open precisely these grounds of challenge. The 1975 Order was different. It contained no statutory peg. It said, in effect, that no one could move any court for the enforcement of Article 21 in any circumstance for the duration of the Emergency.³

The majority's reading was this. If Article 21 is the source of the right to life and personal liberty, and Article 21 has been removed by Presidential Order from the catalogue of provisions a person may move a court to enforce, then there is no source of any right to life and personal liberty that a court may enforce against the State for the duration of the Order. An illegal detention is a detention without authority of law; but if Article 21 is the only constitutional ground for saying that life and liberty may not be taken without authority of law, and Article 21 has been removed, then the constitutional ground is gone. There remain the limits of the statute itself — the MISA — but those limits were among the things the Order forbade the courts from enforcing through habeas corpus.

Beg J put this reading in its clearest form when he replied to Khanna J: 'It seems to me to be legally quite impossible to successfully appeal to some spirit of the Constitution or to any law anterior to or supposed to lie behind the Constitution to frustrate the objects of the express provisions of the Constitution.'⁷ Bhagwati J put it in the register of judicial conscience: 'I do not think it would be right for me to allow my love of personal liberty to cloud my vision or to persuade me to place on the relevant provision of the constitution a construction which its language cannot reasonably bear … the Constitution is the law of all laws and there alone judicial conscience must find its ultimate support, and its final resting place.'⁸

What the four judges were saying — and what is now very hard for us to hear — is that the constitutional text in front of them, drafted by the framers in 1949, amended through Article 359 to permit a Presidential Order of precisely the kind issued, gave the executive what the majority said it gave. They were not lying about this. The Presidential Order said what they said it said. Article 359 said what they said it said. The constitutional text said what they said it said. Their disgrace is not a disgrace of factual error. It is the disgrace of accepting, on the face of the text, a power that should never have been writable into a constitution in the first place — and of refusing to find, in the rule of law itself, a check that the text did not announce.

This is what Khanna J would not accept. He did not pretend the text said something it did not. He argued, instead, that the rule of law contained a postulate prior to the constitutional text — that the State has no power to deprive a person of life or liberty without the authority of law — and that this postulate survived any procedural suspension of Article 21 because it was the condition of constitutional order, not a creature of it.⁶ Beg J's reply was the natural textualist reply: that to appeal to a brooding omnipotence behind the text is to substitute the judge's preference for the constitutional document. It is not a dishonest reply. It is the reply of someone who has decided that the constitutional document, however inconvenient, has to do the work.

Why this matters: honesty about what Article 359 authorised

The reason the majority's honesty matters is that it tells us what the Constitution of 1949, on its own terms, permitted. It permitted Article 359. It permitted a Presidential Order under Article 359 to be unconditional. It permitted that order to remove Article 21 from the catalogue of enforceable rights for the duration of an Emergency that the same executive could declare and continue. The framers wrote this into the document. Not under the duress of 1975 but in the deliberation of 1949. The framers who included Article 22(3), exempting preventive detention from the procedural guarantees of Article 22(1) and (2), and who included Article 359 with its capacity for unconditional suspension orders, knew what they were authorising.¹⁵

The argument that Shivakant Shukla is best read as a moral failure of the four judges flatters the framers. It locates the disgrace at the level of judicial weakness while leaving the constitutional architecture clean. The architecture is not clean. The framers wrote Article 359. They wrote Article 22(3). They wrote an emergency mechanism that permitted the executive both to declare the emergency and to suspend the enforceability of the fundamental rights during it. The four judges of Shivakant Shukla read those provisions and gave them their face value. The country has spent fifty years pretending that this was an aberration of reading rather than what the provisions were always capable of authorising.

There is a precise way to see this. In June 1975 the Supreme Court, in Indira Nehru Gandhi v Raj Narain, struck down a constitutional amendment that sought to insulate the Prime Minister's election from judicial review.¹⁰ The basic-structure doctrine was deployed against the 39th Amendment with full force. The same Court, six months later, surrendered habeas corpus in Shivakant Shukla. The difference is not that the judges had suddenly lost their nerve. The difference is that in Raj Narain the textual battle was over a constitutional amendment, where the basic-structure framework permitted the Court to deploy a doctrine that operated against the text. In Shivakant Shukla, the textual battle was over a Presidential Order issued under Article 359 — a provision of the original Constitution, never struck down on basic-structure grounds, in fact never seriously challenged on such grounds at all. The Court was not invited to declare Article 359 void. It was invited to read a Presidential Order under it in a way the text did not support. Khanna J accepted the invitation. The majority did not. The framers' Article 359 is the part of this story that almost never gets named.

The 44th Amendment as consolation

The constitutional repair came in 1978. The Janata Parliament passed the 44th Amendment, and on the issue of Shivakant Shukla it did three specific things. It added a proviso to Article 359(1) excluding Articles 20 and 21 from any future Presidential Order under that Article. It substituted 'armed rebellion' for 'internal disturbance' in Article 352(1). It rewrote Article 352(3) to require the written advice of the Union Cabinet before the President could proclaim an emergency.⁴ Each of these is real. Each was hard-won. None of them is small. I do not want to be misread as saying that the Janata Parliament's work was symbolic. DD Basu's commentary on Article 359 puts the structural point in plain terms: the 44th Amendment 'cut the Gordian knot' by prohibiting the inclusion of Article 21 in any future Presidential Order.¹¹

I do want to say that the work was incomplete in a structural sense the country has not yet faced. The 44th Amendment took Article 21 off the menu of suspendable rights. It did not take Article 359 off the menu of presidential powers. The provision survives. The Articles other than 20 and 21 remain reachable by Presidential Order — including the whole of Article 19, including Article 14, including the rights of religious minorities under Article 25 and 26. The conceptual structure of Article 359 — that fundamental rights are revocable as a class by executive order during an emergency — was preserved. What changed was the list of which rights are revocable. The list got shorter. The mechanism remained.

The 44th Amendment also did not touch Article 22(3). The carve-out from criminal-procedure guarantees that the framers wrote into the Bill of Rights itself in 1949, allowing the State to detain a person under preventive-detention laws without the safeguards available to a person charged with an actual offence, was left in place. The Maintenance of Internal Security Act, which had been used to detain thousands during the Emergency, was repealed; the National Security Act, with substantially similar architecture, was enacted within thirty months by the same Congress that had returned to power in 1980. The constitutional space for preventive detention without trial was preserved and re-filled. The 1976 judgment was about the suspension of habeas corpus against the executive's use of MISA. The constitutional structure that made MISA possible, and that has made the NSA and the UAPA possible since, was not what the 44th Amendment was directed against. The 44th Amendment was directed against the supplementary mechanism that had permitted MISA's detention orders to be sealed off from any judicial scrutiny at all. Even that mechanism, in attenuated form, remains available against rights other than 20 and 21.

I do not say this to disparage the 44th Amendment. I say it to insist on the gap the Amendment did not close. The Janata Parliament was a reaction to the proclamation of June 1975 and to the judicial surrender of April 1976. It was not a reaction to the framers' Article 359 itself. The constitutional document of 1949 has been amended at the margins of Article 359 without being asked to give up the architecture that made Article 359 thinkable in the first place. The architecture is that fundamental rights, in India, are a class of constitutional protections that the executive can collectively suspend by order during a declared emergency. That this remains true — for rights other than 20 and 21 — is not what the country celebrates when it celebrates the 44th Amendment. But it is what the 44th Amendment left undone.

Maneka, Puttaswamy, and the doctrinal apology

The doctrinal repair came earlier and from a different direction. In January 1978, before the 44th Amendment, the Supreme Court delivered Maneka Gandhi v Union of India and read into Article 21 the requirement of 'fair, just and reasonable' procedure.⁹ The judgment is widely treated as the moment Indian fundamental-rights jurisprudence came of age. It is on every law-school syllabus. It is cited approvingly in every subsequent expansion of Article 21. The settled view treats Maneka as the doctrinal cleansing of Shivakant Shukla.

There is something off about that account, and it is named in the source material. The seven-judge bench in Maneka included Justice P.N. Bhagwati, whose opinion in Shivakant Shukla, twenty months earlier, had described judicial conscience as having to find its 'final resting place' in the constitutional text and had refused to read a substantive content into Article 21 that the text did not announce.⁸ The same judge, in Maneka, read precisely such a substantive content into Article 21 and treated the question as a doctrinal advance rather than a doctrinal reversal. Anup Surendranath has put the awkwardness sharply: Maneka's great virtue was that it raised the level of protection for the freedoms in Article 21 above what Gopalan had set. Its defect was that the bench, including the judge who in Shivakant Shukla had read the rights down to the text and below, treated the new doctrine as if it had always been latent in the same text.¹²

What I want to say about Maneka is that it solved a doctrinal problem without solving an institutional one. The judges in Maneka did not, by reading 'fair, just and reasonable' into Article 21, explain how their reading of the same text in Shivakant Shukla had been compatible with the rule of law they were now reading in. They simply moved on. The Court that delivered Maneka was the same Court that had delivered Shivakant Shukla. It did not, then or later, sit to account for the earlier judgment. The doctrinal U-turn was complete. The institutional reflection was not. The country has accepted the U-turn and accepted the absence of the reflection together, as if the first stood in for the second. It does not.

The formal repudiation came in 2017. In Puttaswamy, a nine-judge bench of the Supreme Court, on the question of the constitutional status of the right to privacy, formally overruled Shivakant Shukla.⁵ D.Y. Chandrachud J, writing for himself and three others, said in terms that the judgment was no longer good law. The overruling was joined in substance by every other opinion on the bench. The judge who delivered the lead overruling was the son of Y.V. Chandrachud J, one of the four judges of the 1976 majority. The symbolism could not have been more public.

I am not against the overruling. The overruling was right. It was overdue. It was the only formal way the Court had of completing the doctrinal repair that Maneka had begun and that the 44th Amendment had constitutionalised. What I am against is the use of the overruling as a closing of the file. Puttaswamy overruled Shivakant Shukla in 2017. It did so on a bench none of whose members had been on the 1976 Court. It did so forty-one years after the judgment had been delivered. It did so in the course of a case about privacy, not in the course of a case about preventive detention or executive supremacy. The overruling came at a distance — temporal, personal, doctrinal — from the conditions that had produced the 1976 judgment. That distance is what made the overruling possible. The same distance is what makes the overruling, by itself, an inadequate answer.

Doctrinal repair, when it comes from a different Court, on a different subject, decades after the event, in a case nobody had selected for the purpose of confronting the original wrong, is real but not sufficient. It is real because the law has to be tidy and the case had to be overruled. It is not sufficient because the country has used the overruling to mark the matter as closed when the structural questions Shivakant Shukla raised — about Article 359, about Article 22(3), about preventive detention without trial, about the framers' tolerance for executive suspension of rights — have not been answered by the overruling. The case was overruled. The architecture was not.

The counter-argument, addressed honestly

The strongest reply to all this is the one that does not concede the central move. It runs as follows. To call Shivakant Shukla 'honest' is to flatter the four judges in a way that does not honour the victims of the Emergency, and to disparage Khanna J in a way that does not honour the only judge on the bench who recognised what was at stake. Khanna J read the same text and found it consistent with the proposition that the State could not detain a citizen without legal authority. The text did not require the majority's reading. The dissent shows it did not. The judgment was not honest. It was wrong about the text, wrong about the rule of law, and wrong about the role of the Court. The settled view's claim that the judgment was a disgrace is the right claim, and the disgrace is precisely that the four judges read into the text a surrender that the text did not require.

This is the strongest version of the reply, and I take it seriously. It rests on Khanna J's dissent, which is the strongest version of the answer to the majority. Khanna J's reading, on the proposition that the State has no power to deprive a person of life or liberty without legal authority, did not require the Court to read into Article 21 a substantive natural-law content that the text did not announce. It required the Court to recognise that the rule of law is the condition of constitutional order, and that the suspension of the enforceability of Article 21 did not by itself confer on the State a positive power to act outside the law. The dissent is, in my view, correct. I have said so. The polemic of this essay is not that Khanna J was wrong. It is that Khanna J's solution required the Court to do something the majority refused to do — to import a postulate that the text did not announce. Khanna J was right to do this. The majority was wrong to refuse. But the refusal was, on its own terms, textually intelligible.

This is the distinction the settled view collapses. The settled view says: the majority was both textually wrong and morally wrong. My polemic says: the majority was morally wrong but textually intelligible, and treating the moral wrong as if it were a textual error allows the country to leave the framers' Article 359 in place. The reply is right that Khanna J showed a different reading was available. My response is that Khanna J's different reading required a move that operates against the text — a move the Court has occasionally made through the basic-structure doctrine, but did not make in Shivakant Shukla and could not have made without confronting Article 359 head-on. That confrontation has still not happened. The reply that 'the text did not require the majority's reading' is, in some sense, true. The reply that 'the constitutional architecture did not produce the majority's reading' is not true. The architecture produced it, the majority accepted what the architecture produced, and the country has spent fifty years dismantling everything except the architecture.

What changes if the polemic is right

Three things change. The first is in how we teach the judgment. The standard pedagogical move is to treat Shivakant Shukla as a case study in judicial cowardice and Khanna J's dissent as the heroic counter-example. The first half of the move is correct. The second half flattens the dissent into iconography. Khanna J was not heroic because he showed the text was clear; he was heroic because he was willing to do what the majority would not — to find, in the rule of law itself, a check that the constitutional document did not announce. To teach the dissent properly is to teach that the rule of law sometimes requires courts to act against, or outside, what the constitutional text on its face authorises. This is a serious doctrinal claim. It has implications for basic-structure adjudication, for the reading of emergency provisions, for the interpretation of Article 22 itself. The country has been celebrating Khanna J's outcome for fifty years while refusing to teach his method.

The second thing that changes is in how we read the 44th Amendment. The Amendment is widely treated as the constitutional bookend of the Emergency. On the reading I have offered, the Amendment is a partial repair of a structural defect that the framers built into the document. The defect is Article 359 read with the catalogue of suspendable rights, and Article 22(3) read with the carve-out from criminal-procedure guarantees. The Amendment narrowed the catalogue but did not retire the mechanism. The structural question — whether a constitutional democracy should write into its document a provision permitting the executive to suspend by order the enforceability of fundamental rights — was never asked. To ask it now would require a constitutional amendment more searching than the 44th. The country has not been willing to ask it. The polemic of this essay is that the unwillingness is itself a residue of Shivakant Shukla. The judgment is read as a one-time disgrace, repaired by partial constitutional reform, and the architecture that produced it is allowed to stand.

The third thing that changes is in how we treat the Puttaswamy overruling. The overruling is celebrated as the doctrinal closure of the Emergency-era constitutional pathology. On the reading I have offered, the overruling is closure of the case but not of the architecture. The 2017 bench, none of whose members had been on the 1976 Court, in a privacy case nobody had selected for the purpose of confronting Shivakant Shukla, formally retired a precedent that had been doctrinally dead since Maneka. The retirement was correct. It was also, in part, ceremonial. The constitutional questions Shivakant Shukla raised about Article 359, Article 22(3), and the institutional vulnerability of the Court remained where they had been. Puttaswamy did not ask them. It overruled the precedent and moved on.

The New York Times editorial of 30 April 1976 said that Justice Khanna's dissent was the only honourable response of the Indian Supreme Court to the Emergency.¹⁴ The editorial added that Khanna J had probably written himself out of the office of Chief Justice. The prediction came true within nine months. Beg J was elevated over him in January 1977; he resigned shortly thereafter; he never sat on the bench again. The country has remembered him as the judge who stood alone, and the memory is correct. What the memory does not do is to honour his method.

Khanna J's method was that the rule of law contains commitments prior to the constitutional text — that the State has no power to deprive a person of life or liberty without legal authority, that an executive officer acting outside the law is acting outside the constitutional order, that a constitutional document cannot, by procedural suspension of a Bill of Rights, confer on the executive a power that the rule of law denies it. This method is not textualism. It is not even basic-structure doctrine in the form the Court has developed it. It is closer to the natural-law commitment that the Indian Court has been reluctant to claim openly, preferring to fold the same commitment into ever wider readings of Article 21. That folding is part of what Maneka began. It is doctrinally productive. It is also evasive. It treats the natural-law commitment as a creature of Article 21 rather than as the condition of Article 21. Khanna J would not have made that move. He thought, correctly, that the rule of law was prior to the constitutional document and survived its procedural suspensions.

To honour Khanna J's method, the country would have to be willing to do two things it has not been willing to do. It would have to teach, in the law schools and in the courts, that the rule of law is not exhausted by what the constitutional document on its face authorises. It would have to confront, at the level of constitutional architecture, the provisions of the 1949 document — Article 359, Article 22(3), the framers' tolerance for executive suspension of rights — that Shivakant Shukla made operative. The first task is intellectual. The second is political. Neither has been done. The fiftieth anniversary of the judgment is, on the polemic of this essay, an occasion not to repeat the disgrace formula but to ask what the doctrine and architecture of the judgment have left intact.

The judgment was the most honest thing the Indian Supreme Court has ever said about what Article 359 authorised. The majority did not lie about the text. They did not invent a power that the framers had withheld. They accepted, with full literalness and at high cost, what the framers had written. The dissent was the only intellectually serious response. The 44th Amendment was a partial structural repair that left the framers' architecture standing. Maneka was a doctrinal advance unaccompanied by an institutional reckoning. Puttaswamy was a formal overruling at a distance of forty years and three benches removed. The country has used these repairs to mark the matter as closed. The matter is not closed. The architecture that permitted four judges to read the text honestly and to surrender habeas corpus by doing so is still in the Constitution. The fiftieth anniversary of ADM Jabalpur is the occasion to say so plainly.

Get the week's biggest decisions, distilled.

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