The 42nd Amendment Was the Constitution

The settled view is that the 42nd Amendment was an Emergency aberration, that the 44th Amendment cleaned it up in 1978, and that the Constitution we now read is essentially the Constitution of 1950. This is the textbook account. It is the account our judiciary aspirants memorise, the account our prelims questions assume, the account that grounds the polite distinction one is supposed to draw between the document the framers gave us and the temporary disfigurement Indira Gandhi imposed. The account is wrong. Most of what we treat as load-bearing constitutional text was inserted in November 1976. The Preamble's two most quoted words. The whole of Part IVA. The textual anchor for legal aid. The first half of Article 31C, which survives. Even the basic-structure doctrine, in the form we now use it, was elaborated principally to police that one amendment. The 44th Amendment was a partial repair, not a restoration. The Constitution we cite is the post-1976 Constitution.

The claim of this essay is unwelcome because it requires giving up a comfort. The comfort is that what is good in our constitutional architecture comes from 1949 and what was bad in it came from 1976. The reality is that some of the best-loved features — the legal-aid clause that runs through Hussainara Khatoon, the duty to protect the environment that runs through MC Mehta, the Preamble word that anchors Bommai's secularism doctrine — are 1976 insertions. They were inserted by a government that suspended elections, jailed the opposition, and censored the press. They were inserted alongside provisions, in the same Amendment Act, that tried to immunise the Prime Minister's election from judicial review and ouster the courts altogether from reviewing constitutional amendments. We discarded the worst clauses and kept the rest. That is a defensible thing to do. It is not the same as pretending the rest never came from where it came from.

The settled view, in its strongest form

Let me put the settled view at its highest, because that is the version this essay disagrees with. On the settled view, the Indian Constitution of 1950 was a substantial, complete document. It contained a Preamble that sufficiently expressed its philosophical orientation. It contained Parts III and IV which between them set out rights and developmental aspirations. It contained Article 368, an amendment power, sufficient to allow change over time. The Emergency package of 1976 disturbed some of that, the Janata package of 1978 repaired most of the disturbance, and what remained — the secular and socialist preambular words, the chapter on duties, the article on legal aid — was either anodyne or beneficial, and certainly compatible with the original architecture. On this view, asking whether the post-1976 Constitution is really a different Constitution is pedantic. The text has grown; the basic project has not changed.

The strongest version of this view is honest. It accepts that the 42nd Amendment was a single legislative package containing both authoritarian provisions and developmental ones, and it accepts that the constitutional system kept some and discarded others. It argues that this is what constitutional systems do: they evolve, they take what they can use, they reject what does not fit. It points out that the basic-structure doctrine itself was a judicial response — a successful one — to the most dangerous clauses of the package, and that this is a vindication of the Indian system rather than an indictment of it. On this account, the post-1976 Constitution is the Constitution functioning as intended, with the courts and Parliament between them filtering what was offered.

Two things make this view comfortable. First, it spares us the embarrassment of saying that any of the document's deeply-valued features came from an Emergency Parliament that no longer had a meaningful opposition.¹ Second, it lets us continue to cite the Constitution as though we were citing what the framers wrote, when in many of the most-quoted respects we are not. The view is comfortable. It is also not quite true. To see why, one has to walk through what the 42nd Amendment actually inserted, what the 44th Amendment chose not to remove, and what the doctrine since 1980 has done with the resulting text.

What 1976 inserted, line by line

The 42nd Amendment ran to fifty-nine sections and amended fifty-three Articles. Its scope was, in DD Basu's measured description, a 'comprehensive review' of the Constitution. The provisions that survive — that were not undone by the 44th Amendment, and have since become canonical — fall into four categories.⁴

First, the Preamble. Two words were added: 'socialist' and 'secular'. A third change inserted the word 'integrity' alongside 'unity' in the phrase 'unity of the nation'. None of this was a small drafting tidying. The Constituent Assembly had specifically considered and rejected an amendment by KT Shah to insert 'socialist' into the Preamble in November 1948. Ambedkar's recorded objection was that to write a particular economic theory into a constitutional text would be to constrain the choices of future generations.¹⁴ In 1949 the choice was deliberate. In 1976 it was reversed by a Parliament that had extended its own tenure by amending Article 83 in the same package. The word 'secular' has a less dramatic Assembly history but its absence from 1949 to 1976 was also deliberate — the framers chose to articulate the substance through Articles 25 to 28 and to leave the noun out of the Preamble. The settled view treats both insertions as cosmetic. They are not. The first half of Bommai turns on the word 'secular' being in the Preamble.¹⁶ Take it out and the basic-structure-of-secularism doctrine has nothing to anchor on; it would have to be derived from substantive Articles, which it can be done, but which is a different and harder doctrinal move.

Second, Article 31C in its original form had been a savings clause limited to laws giving effect to Articles 39(b) and (c). The 42nd Amendment, in section 4, expanded it to cover laws giving effect to all the Directive Principles in Part IV. The change was huge. The original Article 31C, even in Kesavananda Bharati, had been controversial; its second leg, which made the legislature's own declaration conclusive, was struck down at paragraph 1486 of Khanna J's opinion, but its first leg — protection from Article 14, 19 and 31 challenge for Article 39(b)-(c) laws — was upheld.⁵ The 42nd Amendment took that limited carve-out and expanded it to cover the whole of Part IV. Any law that the legislature could plausibly describe as furthering any Directive Principle — a near-unlimited category — would be immune from challenge under the principal rights articles. The 44th Amendment did not roll this back. Minerva Mills did. In 1980, Chandrachud CJ for the majority struck down section 4 of the 42nd Amendment on the ground that the Article 14 and Article 19 rights are themselves part of the basic structure, and the balance between Parts III and IV is itself basic.³ What remains of Article 31C is the original Article 39(b)-(c) carve-out, as upheld in Kesavananda. The text we now read in Article 31C is text the 42nd Amendment put there, partly struck down, partly survived.

Third, Part IVA. The whole chapter on Fundamental Duties is a 1976 insertion. Article 51A was added by section 11 of the 42nd Amendment, on the recommendation of the Swaran Singh Committee.⁸ Ten duties — eleven after the 86th Amendment added clause (k) in 2002 — were placed into the Constitution. The standard textbook formulation is that these duties are non-justiciable, parallel to the Directive Principles, and therefore harmless. This is not quite right. From the early 2000s onwards the Supreme Court has used Article 51A as live constitutional text to read into other doctrines.⁹ The duty to develop scientific temper and humanism in Article 51A(h), the duty to protect the environment in Article 51A(g), the duty to promote harmony in Article 51A(e) — all have done doctrinal work in cases the Court has decided. State of Gujarat v Mirzapur Moti Kureshi Kassab Jamat used Article 51A(g) to uphold an absolute ban on bovine slaughter against an Article 19(1)(g) challenge. Animal Welfare Board of India v A Nagaraja used Articles 51A(g) and (h) to read animal-rights principles into the Prevention of Cruelty to Animals Act 1960.¹⁰ In each case the inserted duty did real work — it shifted the burden, it gave the Court a textual anchor for a value-judgment that would have been harder to make from Part III alone. The duties are not enforceable directly by writ. They are enforceable indirectly through everything the Court does that touches them. That is a meaningful constitutional presence, and it came in 1976.

Fourth, Article 39A. Free legal aid as a constitutional commitment is a 1976 insertion. The article reads: 'The State shall secure that the operation of the legal system promotes justice, on a basis of equal opportunity, and shall, in particular, provide free legal aid, by suitable legislation or schemes or in any other way, to ensure that opportunities for securing justice are not denied to any citizen by reason of economic or other disabilities.' The Legal Services Authorities Act came in 1987, eleven years later. The textual anchor for Hussainara Khatoon's reading of a right to free legal aid into Article 21 is Article 39A.¹¹¹² Take Article 39A away and Hussainara has to derive the right from Article 21 alone, which can be done, but which then makes the right substantially harder to defend against the inevitable argument that the framers in 1949 chose not to write a positive welfare entitlement of that kind into the rights chapter. The 1976 insertion does the doctrinal heavy lifting. The Court relies on it. It is now treated as though it had always been there.

What 1976 attempted, and 1978 reversed

The provisions of the 42nd Amendment that did not survive are well documented. They are also less well understood, because they are usually invoked to make the opposite of the point this essay is making — namely, that the system worked, that the bad clauses were rolled back, and that what remained is good. The provisions themselves are worth listing.

Section 55 of the 42nd Amendment inserted clauses (4) and (5) into Article 368.⁷ Clause (4) provided that no constitutional amendment 'shall be called in question in any court on any ground'. Clause (5) declared that 'there shall be no limitation whatever on the constituent power of Parliament to amend by way of addition, variation or repeal the provisions of this Constitution under this article'. The targets were specifically Kesavananda Bharati and Indira Gandhi v Raj Narain — the two basic-structure judgments delivered in the three years before. The legislative intention was open. Take away judicial review of constitutional amendments and the basic-structure doctrine has no enforcement mechanism. Minerva Mills struck both clauses down in 1980.

Article 31D, inserted by section 6 of the 42nd Amendment, immunised laws against 'anti-national activities' and 'anti-national associations' from Articles 14, 19 and 31. It was repealed by section 4 of the 43rd Amendment in 1977 before the 44th Amendment was passed. Article 32A, inserted by section 14 of the 42nd Amendment, prevented the Supreme Court from considering the constitutionality of State laws under Article 32. Article 226 was substantially curtailed by section 38 of the 42nd Amendment — the High Courts' writ jurisdiction over fundamental rights was retained, but their jurisdiction over other legal rights was conditioned. Both Article 32A and the Article 226 amendments were reversed by the 43rd and 44th Amendments. Articles 105 and 194 were amended to remove judicial review of legislative privileges. Article 357 was amended to give the Centre permanent legislative power over a State after President's Rule. These provisions are gone. The system worked, in the sense that the most egregious clauses were taken out.

What the 44th Amendment chose not to remove tells us something. It did not remove the Preamble insertions. It did not remove Part IVA. It did not remove Article 39A. It did not remove the part of Article 31C that the Court had not struck down. It removed the right to property from Part III (a separate and complicated story which I will not pursue here, except to say that this is itself a substantial structural change). It tightened the Emergency declaration procedure under Article 352. It made Articles 20 and 21 absolutely non-suspendable under Article 359. These were repairs to the structure of executive emergency power. They were not repairs to the substantive constitutional architecture of duties, secularism, and legal aid that the 42nd had laid down.²

The 44th Amendment was passed by a Janata Party government that had every political reason to undo Indira Gandhi's constitutional handiwork in toto. It chose not to. It chose to keep what it considered defensible. That choice is the founding moment of the post-1976 Constitution. From 1978 onwards, what is on the page is what the Janata Parliament chose to keep — which is most of the 42nd Amendment minus the authoritarian clauses. We have been reading that document, and only that document, for forty-eight years.

The basic-structure doctrine grew up policing the 42nd Amendment

The basic-structure doctrine is usually narrated as a story of judicial courage stretching from Sankari Prasad in 1951 through Sajjan Singh, Golak Nath, Kesavananda Bharati, and onward to Minerva Mills and beyond. The narration is correct on the timeline. It misses what the doctrine was actually used for.

Between Kesavananda Bharati in April 1973 and the imposition of the Emergency in June 1975, the basic-structure doctrine was a single 7:6 majority opinion of uncertain force. It had been pronounced; it had not yet done anything. Indira Gandhi v Raj Narain in November 1975 was its first application — striking down Article 329A, which had retrospectively validated the Prime Minister's election.⁶ The basic-structure doctrine in that case had two essential features: judicial review of amendments was permissible, and free and fair elections were part of the basic structure. The Khanna J judgment in Kesavananda had hinted at this; Indira Gandhi v Raj Narain confirmed it.

The 42nd Amendment's section 55 was a direct response to the basic-structure doctrine. It tried to write it out of constitutional text. Minerva Mills in 1980 — which struck down section 55 along with the expanded Article 31C — was the doctrine's second major application, and the application that made it stick.³ Without section 55 to strike down, Minerva Mills would have been a different and smaller case. The expanded Article 31C alone could have been challenged on Kesavananda grounds without needing to test the validity of the ouster clause. By writing the ouster clause and the expanded Article 31C into the same package, the 42nd Amendment forced the Court to articulate the basic-structure doctrine as an enforceable, self-protecting constitutional rule rather than a 7:6 obiter from Kesavananda. The doctrine grew its teeth on the 42nd Amendment.

Waman Rao v Union of India in 1981 did something similar for the Ninth Schedule.¹⁵ The 42nd Amendment had added forty-nine new entries to the Ninth Schedule between 1976 and the date Kesavananda was decided. Waman Rao set a cut-off — laws inserted before 24 April 1973 (the date of the Kesavananda judgment) were immune; laws inserted after would be tested against the basic structure. This cut-off was effectively a doctrine constructed against the 42nd Amendment's pattern of using the Ninth Schedule as a constitutional bypass. IR Coelho v State of Tamil Nadu twenty-six years later confirmed the Waman Rao rule. The whole jurisprudence of how the Ninth Schedule is now policed is jurisprudence designed against the 42nd Amendment, even when the actual statutes inserted came from later Parliaments.¹³

The point is not that the basic-structure doctrine is invalid because it was elaborated against the 42nd Amendment. It is that the doctrine cannot be neatly separated from the constitutional document it was elaborated against. The doctrine as we have it — with its specific features of judicial review of amendments, its specific protection of Article 14 and Article 19, its specific Ninth Schedule cut-off — is a doctrine whose shape was determined by what it had to police. The Constitution it now polices is a Constitution that, on most of the contested clauses, came from the same 1976 enactment that produced the clauses the doctrine struck down.

The textbook account — and where it goes wrong

Take any standard textbook used for judicial-service prelims or for UPSC GS-II. The treatment of the 42nd Amendment runs along familiar lines. The 42nd Amendment is described, usually in two paragraphs, as an Emergency-era package that 'sought to make Parliament supreme'. The 44th Amendment is described, in another paragraph, as having 'restored the constitutional balance'. The textbook then moves on to discuss the Preamble's 'socialist' and 'secular' character — usually in a separate chapter, without flagging that those words came from the same 1976 enactment the textbook just disowned. The Fundamental Duties chapter is treated as part of the original constitutional design, parallel to the Directive Principles. Article 39A is folded into the discussion of legal aid as though it had been there since 1950. The discontinuity is left out.

This is not a malign distortion. It reflects how the system reads itself. The Court does the same. When Bommai cites the word 'secular' from the Preamble to support its basic-structure-of-secularism holding, it does not pause to note that the word came from 1976. When Hussainara cites Article 39A as evidence that the right to legal aid is constitutional, it does not pause to note that Article 39A is a 1976 insertion. When the Mirzapur case cites Article 51A(g) on the environment, it does not pause to note that Article 51A did not exist before 1976. The text is the text; the Court reads it as it finds it. That is how courts work. The text is also a text that did not exist before November 1976. That is how history works.

The textbook account goes wrong in a particular way. It treats the 42nd Amendment as a single package of which some clauses were authoritarian and other clauses were anodyne, and then it discards the authoritarian clauses while quietly absorbing the anodyne ones. The trouble is that 'anodyne' is the wrong description of the surviving clauses. The Preamble words are doing constitutional work. Part IVA is doing constitutional work. Article 39A is doing constitutional work. The surviving part of Article 31C is doing constitutional work. These are not decorative additions. They are textual anchors for substantial doctrine. The textbook can treat them as harmless only because it does not connect them to the doctrines they anchor. Once the connection is made, the picture changes.

The counter-argument: organic constitutional growth

The strongest objection to this essay's claim is that constitutions grow, that the 42nd Amendment is now forty-nine years old, that it has been worked, interpreted, and partially reversed, and that to call its surviving provisions the 'real' Constitution while calling the 1950 text the 'original' Constitution is a kind of pedantry that no functioning constitutional system would tolerate. The American Constitution we cite is the document amended twenty-seven times. The British constitution we cite is statute-law and convention accumulated across centuries. The Indian Constitution we cite is the document as amended one hundred and six times. Carving out one of those amendments — the 42nd — and saying that what came through it is somehow not 'really' constitutional is to deny the document its life. This is the counter-argument at its strongest, and it has weight.

It has weight because constitutions are living documents, and the 42nd Amendment has been alive for nearly fifty years. By a normal accounting, what has been in the text for forty-nine years is the text. The argument that the 42nd Amendment came from a Parliament without an opposition is true and is also irrelevant to whether the inserted text has constitutional validity now. The 44th Amendment, passed by a non-Emergency Parliament, kept most of the insertions. The Supreme Court, in dozens of judgments, has treated the insertions as ordinary constitutional text. The political system, the universities, the bar, the bench — all read the document as the document. The provenance of any particular phrase becomes, on this view, a historical curiosity rather than a constitutional fact.

I accept that the inserted clauses now have constitutional validity. They cannot be removed except by another constitutional amendment, and any such removal would itself be tested against the basic-structure doctrine — which would, ironically, probably preserve them. The Article 39A legal-aid principle would be hard to remove now without producing a basic-structure challenge from the very doctrine the 42nd Amendment provoked into existence. The Part IVA duties would similarly be hard to remove. Bommai would have to be overruled before secularism could be deleted from the Preamble. The text is set. What I do not accept is that the provenance is a historical curiosity. The provenance matters for two reasons.

The first reason is honesty about what we are doing. When a judicial-aspirant cites the Preamble for the proposition that the Constitution commits India to secularism, the aspirant should know that the noun 'secular' was placed in the Preamble in 1976 after the Constituent Assembly had declined to place it there in 1949. The framers' choice in 1949 was deliberate. The 1976 reversal was also deliberate, by a Parliament with different constraints. Both are now part of the constitutional record. To treat the 1976 version as though it had always been there is to misdescribe the document. The misdescription does not change the legal effect — the word is in the text, and the legal effect flows from the text — but it changes what we understand the system to be doing. A Constitution that twice considered and twice resolved the same question is not the same as a Constitution that decided it once and got on with it.

The second reason is the limits of the basic-structure doctrine itself. The doctrine works by treating certain features of the Constitution as immune from amendment because they constitute the document's identity. The features the doctrine has named — secularism, federalism, separation of powers, judicial review, the rule of law, free and fair elections — are features the doctrine derives partly from the Preamble and partly from the substantive Articles. If a substantial part of the Preamble and a non-trivial part of the substantive Articles came in via the 42nd Amendment, then the doctrine is, to that extent, treating the 42nd Amendment as defining the Constitution's identity. The 1976 Parliament did not, of course, have the power to write the basic structure. What it did have was the power to write text that the Court would later read as defining the basic structure. That is what happened. The Constitution's identity, as the doctrine has constructed it over fifty years, has the 42nd Amendment's markings on it.

Where this leaves us

The post-1976 Constitution is a real object. It contains text that was written in 1949 and text that was written in 1976 and text written at various points in between and afterward. The textbook habit of treating the 1949 layer as the Constitution and the 1976 layer as an aberration produces three particular distortions.

It distorts the Preamble. Two of the three most-cited adjectives in the Preamble — 'socialist' and 'secular' — are 1976 insertions. The third, 'sovereign', is 1949. When a judgment opens by citing the Preamble's commitment to a 'sovereign, socialist, secular, democratic republic', it is citing a mixed instrument. The settled view occludes the mixture.

It distorts the chapter on duties. Part IVA was not a structural feature of the 1950 document. It was inserted on the recommendation of a committee constituted by the Indira Gandhi government, in a package that also sought to subordinate the judiciary. The duties chapter has gone on to do real doctrinal work — animal welfare, environment, secular education, scientific temper. The work is now constitutional. The text is constitutional. The textual chapter, however, is a 1976 insertion that the Janata Parliament chose not to repeal.

It distorts the legal-aid jurisprudence. Hussainara Khatoon and the cases that follow rely on Article 39A. Without Article 39A, the right to legal aid would have to be derived from Article 21 alone, which can be done, but which substantially weakens the textual basis. The right depends on a 1976 insertion. The Legal Services Authorities Act 1987 implementing the right is post-1976 legislation. The whole architecture of legal aid as we now have it is post-1976 architecture.

None of this is a brief against the surviving 42nd Amendment provisions. The case for free legal aid is not weakened by knowing that the textual hook was placed in 1976. The case against bovine slaughter, or for it, is not changed by knowing that the duty in Article 51A(g) was inserted in 1976. The framers got many things right; the 1976 Parliament got some things right too; that an authoritarian government can occasionally legislate a defensible substantive provision is not surprising and does not retrospectively justify the authoritarian government's other acts. What the textbook account does, however, is to insulate the 1976 insertions from their provenance by treating them as part of the original document. The insulation is unearned.

If this essay has a single recommendation, it is that the curriculum stop pretending. Every textbook chapter on the Preamble should record that the words 'socialist' and 'secular' were inserted in 1976 by an Emergency-era Parliament that the Constituent Assembly's recorded position had declined to anticipate. Every chapter on Fundamental Duties should record that Part IVA was a 1976 insertion. Every chapter on legal aid should record that Article 39A came in 1976 and the Legal Services Authorities Act came in 1987. Every chapter on the basic-structure doctrine should record that the doctrine's enforceable shape was produced in cases — Indira Gandhi v Raj Narain, Minerva Mills, Waman Rao — that were responding to the 42nd Amendment. The point is not to disown the surviving provisions. The point is to know what one is reading.

The post-1976 Constitution is the Constitution India has. It is not the Constitution Ambedkar reported on the floor of the Constituent Assembly. The two are continuous in important ways and discontinuous in important ways. The textbook account elides the discontinuity. The cost of that elision is small, until it is not — until one is asked to defend the secular character of the Republic and reaches reflexively for a Preamble word that was added by Indira Gandhi, or until one is asked to ground a right to legal aid and reaches reflexively for an Article that was added in the same package as the clauses that tried to abolish judicial review of constitutional amendments. The settled view's account of the document is not the document. The actual document is more interesting, and more compromised, than the settled view permits. We should read it for what it is.

Get the week's biggest decisions, distilled.

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