The constitutional amendment process in India is described almost everywhere in a single sentence about a two-thirds majority, and that sentence is wrong often enough to be useless. It is wrong because the special majority is two requirements, not one. It is wrong because a large class of changes to the Constitution does not go through Article 368 at all. It is wrong because it omits the states, which hold a veto over an identified set of federal provisions and have exercised it in effect within living memory. And it is wrong because it stops before the stage at which amendments have actually been defeated, which is not the floor of either House but the courtroom.

What follows is built to be operated rather than memorised. By the end of it a reader should be able to take any change to the Constitution, proposed or enacted, and place it in the correct route within a minute, state the arithmetic that route demands, name the point at which the proposal is most likely to die, and say whether the courts have anything to work with if it passes. That is a different skill from knowing that Article 368 exists, and it is the skill that the standard account never quite delivers.

The Constitution of India has been amended more than a hundred times. That figure is usually deployed to prove one of two opposite propositions: that the document is dangerously easy to alter, or that its framers built a living instrument responsive to a changing country. Both readings treat amendment as a single undifferentiated activity, which is the root error. Some amendments require nothing more than an ordinary majority in a thin House on a quiet afternoon. Some require a negotiated coalition across party lines in both Houses and the assent of more than half the state legislatures. One of them cleared every procedural hurdle, secured ratification from sixteen state legislatures, and was still struck down. Another was passed with the correct majorities in both Houses and was later held void in substantial part because nobody had noticed that it needed the states. Amendment in India is not hard or easy. It is hard in specific, identifiable places, and this article is about locating them.

What the amending power was built to solve

Every constituent body faces the same problem at the moment it finishes drafting. A constitution that cannot be changed will be broken; a constitution that can be changed at will is not a constitution but a statute with a grand preamble. The two poles were both visible to the Indian Constituent Assembly and both were rejected. At one end sat the American arrangement, where amendment requires two-thirds of both chambers of Congress and ratification by three-quarters of the states, a bar so high that the document has been altered a handful of times in more than two centuries. At the other sat the British position, where there is no distinct amending procedure because there is no codified constitution to amend and Parliament may alter any constitutional arrangement by ordinary legislation.

The Assembly chose neither and it chose deliberately. The scheme that emerged in Article 368, read with the provisions scattered elsewhere that allow certain changes by ordinary majority, is graded. The difficulty of changing a provision is supposed to track how structural that provision is. Adjusting the boundaries of a state, altering the salary of a judge, or abolishing a second chamber in a state legislature was placed within reach of an ordinary majority. Changing a fundamental right or the structure of Parliament was placed behind a special majority. Changing anything that touches the bargain between the Union and the states was placed behind a special majority plus the consent of half the states. That graded design is the single most important fact about the amending power and it is precisely what a summary reciting “two-thirds majority” destroys.

B R Ambedkar defended the choice in the Assembly on grounds that have aged well. His argument was that the framers were legislating for a country whose social and economic arrangements would have to be rebuilt, that the Assembly itself had not been elected on universal adult franchise, and that it would be presumptuous for such a body to bind a future electorate through an unamendable text. The counter-position, that too flexible a document invites the governing party of the day to redesign the state to suit itself, was also made in the Assembly and was answered by the special majority requirement and the federal proviso rather than by rigidity across the board.

Why did the Constituent Assembly reject a single amendment rule?

Because provisions differ in kind. The Assembly graded the procedure so that the effort required to change a provision would track how structural that provision is, which is why altering a state boundary needs an ordinary majority while altering the division of legislative powers between the Union and the states needs a special majority plus the agreement of half the states.

The design has a consequence that its authors probably intended and that later readers routinely miss. Because the routes are graded, the classification question comes first. Before anyone can say what a proposed change requires, they must know which route it travels. A great deal of confused public argument about the Constitution consists of applying the requirements of one route to a proposal that belongs in another.

How Article 368 is structured and which clauses carry the weight

Article 368 is short. It is also not the article the Constituent Assembly wrote, because the amending power has itself been amended twice, once to expand it and once to attempt to place it beyond review. Reading the provision as it now stands without knowing which parts arrived later makes the judicial history unintelligible.

Clause (1) states that Parliament may in exercise of its constituent power amend by way of addition, variation or repeal any provision of the Constitution in accordance with the procedure laid down in the article. The phrase “constituent power” is doing heavy work and it was not in the original text. It arrived through the Constitution (Twenty-fourth Amendment) Act, 1971, which was the Union government’s direct legislative answer to a judgment that had held Parliament could not abridge fundamental rights. The distinction the clause asserts is between ordinary legislative power, which produces laws, and constituent power, which produces constitutional text. That distinction matters because Article 13 renders void any “law” that takes away or abridges fundamental rights, and the whole question in the early litigation was whether an amendment counted as a law for that purpose.

Clause (2) contains the procedure and it is the operative heart of the article. An amendment is initiated by the introduction of a bill in either House of Parliament. The bill must be passed in each House by a majority of the total membership of that House and by a majority of not less than two-thirds of the members of that House present and voting. It is then presented to the President, who shall give assent, whereupon the Constitution stands amended in accordance with the terms of the bill. The proviso to this clause carries the federal requirement, which is treated in its own section below.

Two features of clause (2) are worth flagging immediately because they are where most descriptions go wrong. The first is that the majority requirement is conjunctive. Two separate thresholds must be satisfied at the same time in the same division, and either can be the binding one depending on attendance. The second is that the word “shall” attached to the President’s assent is not decorative. Before 1971 the clause said the bill “shall be presented to the President for his assent and upon such assent being given” the Constitution would stand amended, which left room for an argument that assent was discretionary. The Twenty-fourth Amendment closed that argument by making assent mandatory. There is no return of the bill for reconsideration, no withholding, and no equivalent of the pocket veto that has been debated in relation to ordinary legislation.

Clause (3) says that nothing in Article 13 applies to any amendment made under the article. It arrived with the same Twenty-fourth Amendment and is the textual counterpart to the constituent power language in clause (1). Together the two clauses were an attempt to legislate a judgment out of existence, and they succeeded in part.

Clauses (4) and (5) are the interesting ones, because they were inserted and then removed by two different institutions. The Constitution (Forty-second Amendment) Act, 1976 added a clause providing that no amendment shall be called in question in any court on any ground, and another declaring that there shall be no limitation whatever on the constituent power of Parliament to amend the Constitution. Both were struck down by the Supreme Court in Minerva Mills v Union of India in 1980. They remain in printed copies of the Constitution with a footnote, which is a small but instructive fact: the text of the Constitution and the operative law of the Constitution are not the same document, and anyone who reads only the former will be wrong about the amending power specifically.

The reader who wants the doctrinal arc in full, from the early decisions through Kesavananda Bharati to the position that now governs, should read the article on the basic structure doctrine and what Kesavananda actually decided, which owns that ground in this series. What matters here is narrower and procedural: the amending power was expanded by amendment in 1971, an attempt was made to place it beyond review in 1976, and that attempt failed in 1980. Article 368 as it operates is therefore a provision with a visible scar.

The three routes of the constitutional amendment process in India

Everything that follows depends on getting one thing right at the outset: there are three routes by which the text of the Constitution can be altered, and only two of them are amendments under Article 368. The third is a set of powers conferred elsewhere in the document, exercised by ordinary majority, and expressly declared not to be an amendment for the purposes of Article 368. Treating it as a weaker version of the same procedure is the most common structural error in general writing on this subject, and it produces a cascade of smaller errors downstream.

The first route is the simple majority route. A cluster of provisions permits Parliament to alter constitutional text by the ordinary legislative process, meaning a majority of members present and voting in each House with the usual quorum requirement, exactly as with any other statute. The admission or establishment of new states and the formation of new states by altering the areas, boundaries or names of existing states falls here, along with the consequential changes such a law makes to the First and Fourth Schedules. The abolition or creation of a legislative council in a state falls here, subject to a resolution of the state assembly. Changes to the administration of scheduled areas and tribal areas under the Fifth and Sixth Schedules fall here. So do citizenship provisions, the quorum in either House, the salaries and allowances of members, and the rules of procedure in Parliament. The crucial point is textual and not merely conceptual: the Constitution says in terms that a law made for the reorganisation of states shall not be deemed to be an amendment for the purposes of Article 368, and equivalent language appears in the provisions on legislative councils and in the Fifth and Sixth Schedules. Parliament exercises ordinary legislative power when it does these things, not constituent power.

The second route is the special majority route under Article 368 without ratification. This is the default for amendments and it covers the overwhelming majority of the amendment record: fundamental rights, directive principles, the structure and powers of Parliament, the Preamble, and every other provision not caught by the proviso. Both Houses must pass the bill by the double threshold, and the President must assent.

The third route is the special majority route plus ratification by the legislatures of not less than one-half of the states. It applies only to the identified federal provisions in the proviso to clause (2), and it is the route that changes the character of the exercise from a parliamentary one into a national one.

The classifier below is the tool. Take any amendment, proposed or enacted, and place it in a row.

Route	What it covers	Arithmetic required	Where it dies	Worked example
Ordinary legislative power outside Article 368	Creation and alteration of states with consequential Schedule changes, creation and abolition of state legislative councils, Fifth and Sixth Schedule administration, citizenship provisions, quorum, salaries and allowances of members, parliamentary procedure	Simple majority of members present and voting in each House, with quorum	Rarely on the floor; usually in the political decision to introduce, and for state reorganisation in the requirement of a presidential reference to the state legislature	A States Reorganisation Act creating a new state and rewriting the First Schedule accordingly
Special majority under Article 368	Fundamental rights, directive principles, the Preamble, the structure and powers of Parliament, and all provisions not listed in the proviso	In each House separately, a majority of the total membership of that House and not less than two-thirds of members present and voting	On the floor of the Rajya Sabha, where there is no joint sitting to break a deadlock, and afterwards in court on basic structure grounds	The Constitution (Forty-second Amendment) Act, 1976, whose Article 368 clauses were later struck down
Special majority plus state ratification	Election of the President, extent of the executive power of the Union and the states, the Supreme Court and the High Courts, the distribution of legislative powers between Union and states, any of the Lists in the Seventh Schedule, the representation of states in Parliament, and Article 368 itself	The same double threshold in each House, followed by resolutions of the legislatures of not less than one-half of the states before the bill goes to the President	In the state legislatures, and in court if ratification was skipped	The Constitution (One Hundred and First Amendment) Act, 2016, which created the goods and services tax framework and was ratified by the states

Which route applies to a given amendment?

Read the proviso list first. If the change touches any provision on it, the third route applies. If it does not, check whether the Constitution expressly permits the change by ordinary law and declares it not to be an amendment. If neither, the second route applies by default. The order of those questions matters, because the proviso overrides everything else.

A caution about the classifier that most accounts omit. A single amendment Act can contain provisions falling in more than one route, and when it does, the higher requirement governs the bill as a whole. An amendment bill that changes a fundamental right and also alters an entry in the Union List does not get to be treated as two bills. It needs the special majority for the whole and ratification because of the part, and if the ratification is not obtained the offending part is vulnerable. That is not an academic scenario. It is close to what happened to the Ninety-seventh Amendment, which is examined below.

The first gate: who may move an amendment bill and how it starts

The introduction stage is the least discussed and the most misunderstood part of the procedure, largely because people assume it works like the introduction of a money bill or a bill under Article 3, both of which carry special requirements. It does not.

An amendment bill may be introduced in either House. There is no requirement that it begin in the Lok Sabha, and several significant amendments have been introduced in the Rajya Sabha. There is no requirement of a prior recommendation or sanction of the President, which distinguishes the amending power sharply from the power to form new states, where a bill may be introduced only on the President’s recommendation and after the President has referred it to the legislature of the affected state for its views. There is no requirement that the government move it. A private member may introduce a constitutional amendment bill, and private members have done so regularly across the life of the Republic on subjects ranging from the right to work to the reservation of seats to the abolition of a particular institution.

What private members have not managed to do is get one passed. The record of private member legislation in India is thin across the board, and it is thinner still for constitutional amendment, where the arithmetic requires the governing party to lend its numbers to a proposal it did not draft. The practical consequence is that a private member’s amendment bill functions as an agenda-setting device rather than as a legislative vehicle. It puts a proposal on the record, occasionally forces a minister to state a position on the floor, and sometimes surfaces years later inside a government bill. That is a real function and it is not the same as a route to enactment.

The bill carries the short title format that becomes familiar as the record lengthens: the Constitution (Amendment) Bill, with a serial number assigned in order of introduction. The number the bill carries is not the number the Act will carry. Amendment bills are numbered in the order they are introduced and amendment Acts are numbered in the order they receive assent, and because bills lapse, fail, or sit for years, the two sequences diverge badly. The bill that became the One Hundred and First Amendment Act was introduced as the One Hundred and Twenty-second Amendment Bill, and an earlier attempt at the same reform was introduced as the One Hundred and Fifteenth Amendment Bill and never became anything at all. Anyone reading parliamentary records who assumes the two numbering systems match will misdate half of what they read.

Every amendment bill carries a statement of objects and reasons, and for anyone reconstructing why a change was made this is the first document to read and the one most often skipped. It states the government’s own account of the problem, and because it is prepared before the parliamentary argument begins it is uncontaminated by the concessions that follow. Where the bill has financial implications it also carries a financial memorandum, and where it delegates power it carries a memorandum on delegated legislation. None of these is a constitutional requirement of Article 368; they are requirements of parliamentary practice, and they apply to amendment bills because they apply to bills generally. Their analytical value is that they fix the government’s stated position at a moment when it has no reason to be defensive, which makes the comparison with the enacted text considerably sharper.

One route is closed at this gate and the closure is absolute. The Constitution cannot be amended by ordinance. The power to promulgate an ordinance when Parliament is not in session produces an instrument with the same force and effect as an Act of Parliament, and an Act of Parliament made in the exercise of ordinary legislative power cannot alter the Constitution. Constituent power is exercisable only by Parliament in the manner Article 368 specifies. Nor can the Constitution be amended by a referendum, because the document contains no provision for one. Proposals to introduce a referendum requirement have been floated periodically and would themselves require an amendment to Article 368, which is one of the provisions in the proviso and therefore needs the states.

The second gate: what the special majority actually requires

Here is the single most repeated error in the field, and it is worth stating plainly before the arithmetic. The special majority is not two-thirds. It is a majority of the total membership of the House and not less than two-thirds of the members present and voting, both at once. A division that clears one and misses the other fails.

Take the two limbs separately. The total membership limb is calculated on the total number of seats the House comprises, not on the number of members who happen to hold those seats on the day. Vacancies do not reduce the denominator. If seats are unfilled because of deaths, resignations, disqualifications or elections not yet held, the threshold stays where it was. This is a fixed, knowable number that does not move with attendance, and it functions as a floor beneath the whole exercise.

The present and voting limb is calculated on the members who actually cast a vote in the division. Members who are absent are not present and voting. Members who are present but abstain are not voting, and an abstention is therefore not a vote against. This limb floats with attendance and it can be satisfied by a surprisingly small number of members in a thin House, which is exactly why the first limb exists.

Work the arithmetic on a Lok Sabha of 543 seats. The total membership limb requires 272 votes in favour, and that number does not change no matter who turns up. The two-thirds limb requires two-thirds of whoever votes. If 400 members vote, two-thirds is 267, which is below 272, so the total membership limb binds and 272 is the real target. If 500 members vote, two-thirds is 334, comfortably above 272, so the two-thirds limb binds. The crossover sits at three-quarters attendance: once about 408 members are present and voting, the two-thirds requirement overtakes the total membership requirement and becomes the operative constraint. Below that line the government is chasing 272; above it, the government is chasing a moving number that rises with every additional member who votes.

That crossover is the most useful single piece of arithmetic in this area and it explains a pattern that otherwise looks strange. A government confident of its numbers wants a full House, because a full House signals consensus. A government scraping through wants the minimum viable House, because a thinner division lowers the two-thirds bar while the total membership bar stays fixed. Neither strategy can help a government that lacks 272 committed votes in the Lok Sabha, because that limb is immovable.

Run the same exercise on a Rajya Sabha of 245 seats. The total membership limb requires 123. Two-thirds overtakes 123 once about 185 members are present and voting. The Rajya Sabha typically sees higher proportional attendance in significant divisions than its larger counterpart, which means the two-thirds limb binds more often in the Upper House than in the Lower.

Does an abstention help or hurt an amendment bill?

An abstention neither helps nor hurts directly, because a member who is present but does not vote is excluded from the present and voting count. It lowers the two-thirds threshold by shrinking the denominator, while leaving the total membership threshold untouched. For a government short of the fixed limb, abstentions are worthless; for a government short only of two-thirds, they are quietly useful.

The requirement operates in each House separately and there is no aggregation. A bill that clears the Lok Sabha by an overwhelming margin and falls three votes short in the Rajya Sabha has failed, and the Lok Sabha margin is irrelevant to that outcome. The same double threshold applies to the bill and, under parliamentary practice, to each clause of it and to any amendment moved to the bill, which is a detail that matters in contested passages where the opposition attempts to strip particular clauses out.

Why the Rajya Sabha is an absolute veto point

Ordinary legislation in India has a deadlock-breaking mechanism. Where the two Houses disagree on a bill, or one House rejects it, or more than six months pass without the second House disposing of it, the President may summon a joint sitting of both Houses, at which the matter is decided by a majority of the total number of members of both Houses present and voting. Because the Lok Sabha is roughly twice the size of the Rajya Sabha, a joint sitting overwhelmingly favours whoever commands the Lower House. That mechanism has been used sparingly, but its existence shapes negotiation: an Upper House that knows it can ultimately be overridden bargains differently from one that knows it cannot.

The joint sitting provision does not apply to constitutional amendment bills. This is not an inference from silence or a matter of convention. The deadlock machinery is drafted for bills passed by one House and rejected or unaddressed by the other in the ordinary legislative process, and Article 368 lays down a self-contained procedure requiring passage in each House by the specified majorities. A joint sitting could not satisfy that requirement, because the requirement is expressed House by House. Money bills are also outside the joint sitting machinery, for a different reason, and the general architecture of the ordinary legislative process is set out in the article on how a bill becomes law in India, which owns that ground and which this article deliberately does not restate.

The consequence is the single most underexplained fact about amending the Indian Constitution. The Rajya Sabha holds an absolute veto over every constitutional amendment. Not a delaying power, not a suspensive veto exercisable for six months, not a power that can be overridden by a joint sitting or by resubmission in the next session. A government that cannot assemble a majority of the total membership of the Rajya Sabha, and two-thirds of those voting there, cannot amend the Constitution, and no size of Lok Sabha majority alters that.

Because the Rajya Sabha is elected indirectly by state legislative assemblies on staggered six-year terms with roughly a third retiring every two years, its composition lags the Lok Sabha by several years. A party that wins a landslide general election does not thereby acquire the Upper House. It acquires the Upper House gradually, through state election victories that feed into Rajya Sabha seats as vacancies arise, and it can lose it the same way. The practical effect is that constitutional amendment in India is structurally a multi-year project for any government that starts without the numbers, and a negotiation with other parties for almost every government that has ever attempted it.

The failed panchayati raj amendments of 1989 are the cleanest illustration. Two constitution amendment bills designed to give constitutional status to local government were pushed by a government with a very large Lok Sabha majority. One was defeated in the Rajya Sabha and the reform stalled, and a subsequent attempt lapsed. The same reform succeeded three years later under a government with a far weaker position in the Lower House, because that government negotiated the text with the states and the opposition rather than relying on numbers. The implementation record of those amendments after they finally passed is examined in the article on what the 73rd and 74th Amendments actually devolved, and the passage story is the standing rebuttal to the idea that a large majority is what amendment requires.

The third gate: the ratification proviso and the federal provisions

The proviso to Article 368(2) is where the Constitution stops being a document about Parliament and becomes a document about a federation. It provides that if an amendment seeks to make a change in an identified set of provisions, the amendment must also be ratified by the legislatures of not less than one-half of the states by resolutions passed to that effect, and that this must happen before the bill is presented to the President for assent.

The identified provisions cover, in substance, the parts of the Constitution that define the relationship between the Union and the states rather than the internal arrangements of either. The election of the President is included, because the electoral college is composed of members of Parliament and of the state legislative assemblies, so the manner of election is a shared property. The extent of the executive power of the Union and of the states is included, because that is the boundary line of administrative authority. The Union judiciary and the High Courts are included, since the courts adjudicate the federal division and their structure cannot be a matter for one side of it alone. The distribution of legislative powers between the Union and the states is included, as is any of the Lists in the Seventh Schedule, which is where that distribution is actually written down. The representation of states in Parliament is included, meaning the allocation of Rajya Sabha seats. And Article 368 itself is included, which is the entrenchment clause: the amending power cannot be redesigned by Parliament alone.

Two structural observations about that list are worth more than a recitation of it. The first is that the list is drawn by subject matter and not by dignity. Fundamental rights are not on it. The Preamble is not on it. Some of the most consequential amendments in Indian history, including those that reshaped the right to property and those that rewrote the free speech provision, did not require a single state to agree. Meanwhile a technical adjustment to an entry in the State List, which no newspaper would report, requires half the states to pass resolutions. The proviso protects federalism, not rights, and readers who expect it to protect what they consider most important will misclassify constantly.

The second observation is that the trigger is whether the amendment “seeks to make any change” in those provisions, which is broader than amending their text directly. An amendment that leaves the words of the Seventh Schedule untouched but effectively transfers a subject out of the reach of state legislatures is making a change in the distribution of legislative powers. This is the interpretive question at the heart of the litigation examined below, and it is the reason a government cannot avoid the proviso by careful drafting alone.

Do all constitutional amendments need approval from the states?

No, and the belief that they do is one of the most widely repeated errors about the Indian Constitution. State ratification is required only where the amendment seeks to change one of the federal provisions listed in the proviso to Article 368(2). Most amendments, including many of the best known, were enacted by Parliament alone with no state involvement at all.

The threshold is one-half of the states and not two-thirds or three-quarters, which is a genuinely low federal bar by comparative standards. It is also a bar with no time limit attached. The Constitution does not specify a period within which the state legislatures must act, which means a bill can sit awaiting the last few ratifications indefinitely, and the government of the day controls the timing of the presentation to the President. Ratification is by resolution of the state legislature and not by any executive act, so a state government cannot ratify by notification and a Governor cannot ratify on the state’s behalf. In a state with two chambers the practice has been for the resolution to go through the assembly, since the constitutional text speaks of the legislature of the state.

The Constitution is silent on several questions that would matter if they ever arose squarely. It does not say whether a state may withdraw a ratification already passed before the bill goes to the President. It does not say what happens if a state ratifies a version of the bill that is subsequently amended in Parliament. It does not say whether union territories with legislatures count towards the tally, though the settled understanding is that the requirement speaks of states. These silences have not produced a crisis, largely because governments seeking ratification have generally sought comfortable margins rather than exactly half, but they are real gaps and a reader should know they exist rather than assume the scheme is airtight.

What ratification looks like in practice

The formal description of ratification makes it sound like a constitutional formality. The record shows something closer to a political campaign, and the difference is instructive.

When the Union government wants a federal amendment, it must persuade half the states to spend legislative time on a resolution that carries no direct benefit to the state passing it and often involves surrendering something. States governed by the same party as the Union are ordinarily reliable. States governed by others are not, and the negotiation with them is where the substance of the amendment usually changes. The most instructive episode in the entire ratification record is the goods and services tax amendment, where the eventual text dropped a proposed additional levy on inter-state supply that states had objected to, guaranteed compensation for revenue loss for a defined period, and built the voting structure of the new council so that neither the Union nor the states could act alone. Those concessions were the price of ratification, and they are visible in the constitutional text itself rather than in a side agreement. The design of that bargain is examined in detail in the article on the 101st Amendment and how the GST Council votes.

The sequencing is worth being precise about because it is frequently reported wrongly. Ratification comes after both Houses have passed the bill and before the President assents. It is not a precondition of introduction and it is not a step that follows enactment. A state that ratifies is ratifying a bill, not an Act, and the resolutions are typically moved by the state government within weeks of parliamentary passage, sometimes in specially convened sittings. Once the required number of resolutions has been communicated, the bill goes forward and the President’s assent completes the amendment.

A further practical point about sequencing is worth recording because it shapes the negotiation. Since ratification follows parliamentary passage, the states are being asked to approve a text that can no longer be altered, and the Union government therefore has an interest in settling state objections long before the bill reaches a vote. The forums in which that settlement happens are political rather than constitutional: meetings of chief ministers, councils of ministers drawn from the states, and the standing machinery of intergovernmental consultation. None of it is required by Article 368 and all of it is functionally necessary, which is a good illustration of how much of the working constitution sits outside the constitutional text. Where such consultation has been thorough, ratification has been quick and largely uncontested. Where it has not, the ratification stage has either stalled or been abandoned before it began, and the proposal has quietly disappeared from the legislative programme without ever being defeated on a vote.

There is no constitutional requirement that the states which ratify be geographically or politically representative, and no requirement that the ratifying resolutions be identical in terms. The requirement is a count. That formalism is a feature rather than a defect, because a substantive test of state consent would be unadministrable, but it does mean that an amendment can technically clear the federal gate with the agreement of half the states while the other half objects loudly, and several have.

The 97th Amendment and the proof that the proviso has teeth

Everything said so far about the ratification requirement could be dismissed as procedural detail if the requirement had never been enforced. It has been, and the case that enforced it is the most useful single anchor in this entire subject because it converts an abstract rule into a demonstrated consequence.

The Constitution (Ninety-seventh Amendment) Act, 2011 dealt with co-operative societies. It added the right to form co-operative societies to the freedoms guaranteed by Article 19(1)(c), inserted a directive principle on their promotion, and, most significantly, added a new Part IXB to the Constitution containing a detailed code governing the incorporation, regulation, board composition, term of office, audit, and elections of co-operative societies. The detail is the point. Part IXB did not set out a principle; it set out a regulatory scheme of the kind ordinarily found in a state statute.

Co-operative societies are a State List subject. By writing a detailed governance code for them into the Constitution, Parliament was doing something to the distribution of legislative powers between the Union and the states, whatever the drafting said on its face. The amendment was passed by both Houses with the special majority. It was not ratified by any state legislature, because the Union government took the view that the proviso was not attracted.

The Gujarat High Court held otherwise and struck down Part IXB. On appeal, the Supreme Court delivered its decision on 20 July 2021 in Union of India v Rajendra N Shah. A three-judge bench held by majority that Part IXB, insofar as it applied to co-operative societies operating within a single state, was void for want of ratification under the proviso to Article 368(2). The reasoning is the part worth carrying away: an amendment that substantially restricts the field of legislation available to state legislatures is a change in the distribution of legislative powers even where it does not alter the text of the Seventh Schedule, and the proviso is engaged by substance rather than by form.

The court then did something that makes the case doubly useful. Rather than striking down the whole of Part IXB, the majority severed it, holding that the Part survives so far as it concerns multi-state co-operative societies, which fall within Union legislative competence and therefore raise no federal objection. One judge on the bench took the view that severance was not available and that the entire Part should fall. That difference is a genuine and unresolved question about how far a court should go in salvaging an unratified amendment, and a reader should know it exists rather than treat the outcome as unanimous.

Three consequences follow and they are the reason this case belongs in a pillar article rather than a footnote. First, the ratification requirement is enforceable and has been enforced, so the classification exercise at the start of this article is not academic. Second, the test is substantive, which means a government cannot escape the proviso by avoiding the Seventh Schedule and legislating around it. Third, an amendment can be partly valid, which is a proposition many readers find surprising and which follows from ordinary severability principles applied to constitutional text.

The framing worth remembering is that the Ninety-seventh Amendment failed at the third gate alone. It cleared introduction. It cleared the special majority in both Houses. It was never seriously attacked on basic structure grounds. It failed because a procedural requirement was not complied with, and procedure in constitutional amendment is not a formality but a distribution of power.

The fourth gate: judicial review and the limit that cannot be legislated away

An amendment that clears introduction, both Houses, ratification where required, and assent is part of the Constitution. It is also, from that moment, capable of being challenged.

The proposition that a constitutional amendment can be judicially reviewed at all is not obvious and was not always the law. The early decisions held that Parliament’s power under Article 368 extended to fundamental rights and that an amendment was not a “law” within the meaning of Article 13, so the ordinary voiding provision did not reach it. That position was reversed in 1967 and then reversed again, in a modified form, in Kesavananda Bharati v State of Kerala in April 1973, where a bench of thirteen judges held by a majority of seven to six that Parliament may amend any part of the Constitution including the chapter on fundamental rights, but may not damage or destroy its basic structure. The Twenty-fourth Amendment, which had asserted constituent power and excluded Article 13, was upheld. The limitation the court substituted was not textual but structural.

Seven years later the court applied that limitation to the amending power itself. In Minerva Mills v Union of India the two clauses that the Forty-second Amendment had inserted into Article 368, purporting to exclude judicial review of amendments and to declare the amending power unlimited, were struck down. The reasoning has a neatness worth stating: a limited amending power is itself part of the basic structure, so an amendment that converts a limited power into an unlimited one destroys the very feature it relies on. Parliament cannot use the amending power to enlarge the amending power beyond its own limits, and the attempt in 1976 to do exactly that is the reason this is settled rather than theoretical. The political context of that attempt, and the surprisingly large portion of the Forty-second Amendment that survives untouched, are the subject of the article on the 42nd Amendment as the Emergency charter.

The doctrine’s content, its case-by-case development, the features that have been recognised as basic and the reasons the list has deliberately never been closed, belong to the judicial interpretation article in this cluster and are not restated here. What belongs here is the procedural consequence for anyone assessing an amendment. The fourth gate is real and it has been decisive. The clearest single instance is the amendment that replaced the collegium system for appointing judges with a commission: it was passed with the required majorities, it was ratified by well over half the state legislatures, and it was struck down in 2015 by a constitution bench holding that it damaged judicial independence. Every procedural requirement was satisfied and the amendment still failed.

That case and the co-operative societies case make a useful pair, because they fail at opposite gates for opposite reasons. One satisfied every substantive expectation and fell on a procedural omission. The other satisfied every procedural requirement and fell on substance. Together they establish that all four gates are live, which is the proposition this article is built around.

The honest qualification is that basic structure review of amendments is exercised sparingly. Amendments are struck down rarely, and most of the amendment record has never been seriously challenged. A reader who comes away thinking the courts routinely veto amendments has taken the wrong lesson. The correct lesson is narrower: the gate exists, it is not a formality, and its existence changes how amendments are drafted, because government law officers now advise on basic structure exposure at the drafting stage rather than after the fact.

The simple majority route and why it is not an amendment under Article 368

The third route deserves its own treatment because collapsing it into Article 368 produces confident, wrong answers to some of the most commonly asked questions about the Constitution.

Start with the strongest example. Parliament may by law admit or establish new states, and may form a new state by separation of territory from any state, by uniting two or more states or parts of states, or by uniting any territory to a part of any state. It may alter the area, boundaries or name of any state. Such a law may contain provisions for amending the First Schedule, which lists the states and their territories, and the Fourth Schedule, which allocates Rajya Sabha seats among them. And the Constitution states expressly that no such law shall be deemed to be an amendment for the purposes of Article 368.

Read that sequence carefully and the significance emerges. Creating a new state changes the text of two Schedules and alters the allocation of seats in the Upper House, which sounds exactly like the sort of federal change that would attract the ratification proviso. It does not, because Parliament is not exercising constituent power. It is exercising an ordinary legislative power conferred elsewhere, subject to a different safeguard: the bill can be introduced only on the President’s recommendation, and the President must first refer it to the legislature of the state whose boundaries or name would be affected, for that legislature to express its views within a specified period. The views are not binding. A state can object and be reorganised anyway, which is a fact about Indian federalism that surprises readers who assume the ratification proviso covers state boundaries. The federal division and the state map are protected by different mechanisms of very different strength.

The pattern repeats elsewhere. A state legislative council may be created or abolished by Parliament by ordinary law, provided the legislative assembly of the state has first passed a resolution to that effect by a special majority of its own. The administration of scheduled areas under the Fifth Schedule and of tribal areas under the Sixth Schedule may be altered by parliamentary law, and both Schedules contain provisions declaring that such a law is not an amendment for the purposes of Article 368. Citizenship provisions are subject to any law made by Parliament. The quorum of each House, the salaries and allowances of members, the rules of procedure, the use of English in Parliament and the courts, and the number of judges of the Supreme Court are all matters the Constitution leaves to ordinary legislation.

What all these have in common is that the Constitution itself contemplated further legislative elaboration. The framers wrote a framework and left the detail to Parliament, and Parliament filling in that detail is not amending the framework even when the exercise changes constitutional text. That is the conceptual distinction, and it does real analytical work. It explains why the reorganisation of states, which has redrawn the political map of India repeatedly, has never once required a special majority. It explains why the number of Supreme Court judges has been increased several times by ordinary Acts that nobody describes as constitutional amendments. And it explains why a question like “can this be done by simple majority” cannot be answered by looking at how important the change feels.

Assent, commencement, and what happens after both Houses pass

The stages after parliamentary passage are short but not empty, and two of them regularly produce confusion.

The President’s assent is mandatory. Since the Twenty-fourth Amendment the clause has read that the bill shall be presented to the President who shall give assent, which removes the discretion available in relation to ordinary bills, where the President may withhold assent or return a bill that is not a money bill for reconsideration. There is no such power here. The President cannot return a constitutional amendment bill, cannot withhold assent, and cannot sit on it indefinitely as a matter of constitutional obligation, though the Constitution prescribes no period within which assent must be given. The office has never tested this, and the practice has been prompt assent.

On assent the Constitution stands amended in accordance with the terms of the bill. That phrase is doing something specific. The amendment takes effect as the bill provides, which means the commencement structure is whatever the amendment Act itself specifies. Some amendments come into force at once on assent. Some specify a future date. Many provide that the Act shall come into force on such date as the Central Government may by notification appoint, and several provide different dates for different provisions. This is where a genuine trap sits for readers and even for careful writers: an amendment can be enacted, printed in the Constitution with a footnote, and not be in force, or be in force only in part. The Ninety-ninth Amendment on judicial appointments was brought into force by notification more than a year after it received assent, and the amendments creating panchayats and municipalities specified commencement dates months after enactment. Anyone stating that a provision is operative should check the commencement notification rather than the assent date, and this is the most common source of confidently wrong statements about what the Constitution requires.

The Act is then numbered in the sequence of amendment Acts by date of assent and carries the short title format the Constitution (Amendment) Act with the year of enactment. Amendments are commonly referred to by their number alone, which works well enough in conversation and badly in research, because the number tells the reader nothing about the subject and because the bill numbers and Act numbers do not correspond.

Where amendment bills actually die

The four gates describe where an amendment can fail as a matter of law. The record shows where they fail as a matter of practice, and the two lists overlap only partly.

The largest single cause of failure is dissolution. A bill pending in the Lok Sabha lapses when that House is dissolved, as does a bill passed by the Lok Sabha and pending in the Rajya Sabha. A bill pending in the Rajya Sabha which has not been passed by the Lok Sabha does not lapse, because the Upper House is a continuing body that is never dissolved. Prorogation does not cause a bill to lapse. Since a constitutional amendment often requires long negotiation with states and opposition parties, the calendar is a real adversary: a reform that needs eighteen months of persuasion and is introduced in the fourth year of a Lok Sabha is racing a clock. The first serious attempt at the goods and services tax amendment lapsed exactly this way, and the reform had to be reintroduced as a fresh bill in the next Parliament.

The second cause is the Rajya Sabha arithmetic already described. The third, and more common than the record of divisions suggests, is anticipated failure. Governments do not usually introduce constitutional amendments they expect to lose, because a defeat on a constitutional bill is a serious political event. Proposals therefore die in cabinet, in consultation with allies, or in the Law Ministry, and the visible record of failed amendment bills understates the number of amendments that were seriously contemplated and abandoned.

The fourth is committee scrutiny, which functions here as both a delay and a quality mechanism. There is no constitutional requirement that a constitutional amendment bill be referred to a committee, which is worth stating plainly because many readers assume otherwise. Referral to a select committee, a joint committee, or a departmentally related standing committee is a matter of parliamentary practice and political negotiation. Significant amendments have gone through detailed committee stages, and significant amendments have been passed within days of introduction. The presence or absence of committee scrutiny is one of the most reliable indicators of how contested a passage was, and it is the first thing to look for when reading the record of any amendment.

The pattern that emerges from all four causes is that constitutional amendment in India is a negotiation rather than an exercise of majority power. The great majority of successful amendments have passed with support from parties outside the government, often including parties that opposed the government on everything else in that session. That is not because Indian politics is unusually consensual. It is because the arithmetic in the two Houses, plus the ratification requirement for federal provisions, makes unilateral amendment arithmetically unavailable to almost every government that has ever held office.

Reading the amendment record by era

More than a hundred amendments across the life of the Republic is a large number, and the number by itself misleads. The record is not a steady drip. It clusters, and the clusters correspond to distinct projects of the state.

The first cluster runs through the first two decades and is dominated by land. The abolition of intermediary tenures and the redistribution of agricultural holdings was the central domestic project of the early Republic, it ran immediately into the property guarantee in the chapter on fundamental rights, and it produced a sequence of amendments designed to insulate land reform statutes from challenge. The device invented for this purpose was the Ninth Schedule, a list of enactments declared immune from fundamental rights challenge, created by the Constitution (First Amendment) Act, 1951 with a small number of entries and expanded steadily thereafter. The passage of that first amendment, by a Parliament that had not been elected and within about a year of the Constitution coming into force, is examined in the article on why the First Amendment came so fast, and it establishes the pattern that dominates this era: an amendment as a reply to a judgment.

The second cluster is the confrontation between Parliament and the courts over the amending power itself, running from the mid-1960s to 1980. This is the period that produced the Twenty-fourth Amendment asserting constituent power, the amendments shielding directive-principle legislation from rights challenge, the Thirty-ninth Amendment placing an election dispute beyond judicial reach, the Forty-second Amendment attempting to abolish review of amendments altogether, and then the reversals of 1977 and 1978. It is the only period in which the amending power was itself the subject of amendment, in both directions.

The third cluster is federal and institutional and begins in the late 1980s. Its characteristic amendment does not adjust a right; it creates an institution or entrenches a level of government. Constitutional status for panchayats and municipalities, the creation of a national commission with constitutional status, the creation of the goods and services tax council, and the constitutional entrenchment of bodies previously created by statute all belong here. These are the amendments most likely to require ratification, because creating a permanent institution with defined powers almost always touches the distribution of legislative competence.

Running alongside all three is a fourth category that receives almost no attention and accounts for a large share of the total: extensions of the reservation of seats for Scheduled Castes and Scheduled Tribes in the legislatures, adjustments to the schedule of official languages, changes consequential on the reorganisation of states, and technical corrections. These pass without controversy, often without a division, and they are the reason the raw count of amendments is a poor proxy for constitutional instability.

Can a government with a large Lok Sabha majority amend the Constitution at will?

No. It must also command a majority of the total membership of the Rajya Sabha and two-thirds of those voting there, with no joint sitting available to break a deadlock. For federal provisions it must additionally obtain resolutions from half the state legislatures, and the amendment remains open to challenge on basic structure grounds after enactment.

The honest counterpoint has to be stated with equal force, because a reader who takes only the reassuring half of this analysis will be as badly misled as one who takes only the alarming half. The Constitution has in fact been amended more than a hundred times, which is many times more often than most codified constitutions of comparable age. Governments have amended it to reverse judgments they disliked, to protect legislation from constitutional challenge, to extend their own terms, and to restructure federal relations. The gates are real, and they have also been passed through repeatedly. The correct conclusion is not that amendment is difficult but that the difficulty is unevenly distributed: trivially easy for the classes of change the Constitution left to ordinary law, moderate for most Article 368 amendments where a government has Upper House numbers or can negotiate them, high for federal provisions, and genuinely uncertain for anything that touches the structural features the courts have identified.

Can the Preamble be amended?

This question recurs perpetually in public argument and it has a clean answer that most of the argument ignores. The Preamble has been amended. The Constitution (Forty-second Amendment) Act, 1976 altered it, inserting two descriptive words and changing a phrase about the unity of the nation. Whatever else is disputed about that amendment, the fact that it reached the Preamble is not.

The legal question underneath is whether the Preamble is part of the Constitution for the purposes of Article 368, and the answer settled in the affirmative in the Kesavananda litigation, displacing an earlier view that the Preamble was an introductory statement outside the operative text. Being part of the Constitution, it is amendable by the ordinary Article 368 route, which is the second route in the classifier, since the Preamble is not among the provisions listed in the ratification proviso.

That is where the legal answer stops and where the more interesting question begins. An amendment to the Preamble is subject to the same basic structure limitation as any other amendment, and the Preamble is the most concentrated statement of the features the courts have drawn on when identifying what is basic. An amendment removing a word from the Preamble would therefore raise a challenge not because the Preamble is unamendable but because the substance being removed may be structural. The distinction between those two propositions is exactly the distinction that public argument on this subject collapses, and holding it firmly is a good test of whether a reader has understood the amending power at all.

The four gates: using the classifier on a live proposal

The analytical claim this article advances is that constitutional amendment in India has four gates, and that an amendment fails if it misses any one of them. Introduction, which is permissive and rarely decisive but which closes off the ordinance and the referendum entirely. The two-part special majority in each House, where the Rajya Sabha holds an absolute veto. The ratification proviso, which is engaged by substance and not by drafting. And judicial review under the basic structure limitation, which operates after enactment and cannot be legislated away.

Applying the four gates to a proposal is a short exercise once the habit is formed. Take any proposed change and ask, in order: is this something the Constitution allows to be done by ordinary law and declares not to be an amendment, in which case none of the gates apply and a simple majority suffices. If not, does the change touch the election of the President, the executive power of the Union or the states, the Supreme Court or the High Courts, the distribution of legislative powers, a List in the Seventh Schedule, the representation of states in Parliament, or Article 368, either directly or by substantially restricting what state legislatures may do, in which case ratification is required and the government needs a state strategy before it needs a floor strategy. Then ask what the arithmetic looks like in the smaller House, because that is where the proposal will live or die. Then ask what a challenge would look like if it passed, which means asking whether the change removes a structural feature rather than adjusting a policy.

Run that sequence on the co-operative societies amendment and the failure is visible in advance. The change was detailed regulation of a State List subject, which engages the distribution of legislative powers by substance; the government did not seek ratification; the amendment was therefore exposed at the third gate from the day it was introduced, and the exposure had nothing to do with whether the policy was sound. Run it on the judicial appointments amendment and the exposure is at the fourth gate instead, since the change altered the mechanism by which judges are appointed and judicial independence is among the features the courts have treated as structural. Run it on the goods and services tax amendment and the exposure is at the third gate too, but the government identified it correctly, obtained the ratifications, and passed.

The point of the exercise is not prediction, which is unavailable in constitutional litigation. It is diagnosis. A reader who can say which gate a proposal is most exposed at, and why, understands the amending power better than a reader who can recite Article 368 from memory.

The arguments that are actually being had

Three claims circulate constantly whenever constitutional amendment becomes a live political subject, and each deserves a direct answer rather than a diplomatic one.

The first is that a government with a big enough mandate can rewrite the Constitution overnight. The arithmetic answers most of this. Rewriting the Constitution would require sustained majorities of a specific kind in both Houses, with the Upper House composition lagging by years and being replaced in staggered tranches, plus ratification by half the states for any federal provision, plus survival of basic structure review for anything structural. What the claim gets right, and what the reassuring answer often glides over, is that a determined government with sustained numbers can accomplish a great deal within those constraints, as the record between 1971 and 1976 demonstrates. The constraint is not that constitutional change is impossible; it is that it is slow, visible, and reversible, and those three properties are what the design actually buys.

The second is that the basic structure doctrine is judicial overreach with no textual foundation. This is a serious position held by serious people and it should not be dismissed. The words “basic structure” appear nowhere in the Constitution. The doctrine was announced by a bench divided seven to six, in a judgment whose summary of the majority position was itself contested at the time. It confers on unelected judges the power to void the considered act of a constituent body, and its content has been developed case by case with no closed list, which means the limits on Parliament’s power are defined by the institution that enforces them. The counter-case is equally serious: an amending power without limits permits constitutional self-destruction by ordinary political means, every constitutional system resolves that problem somehow, and the alternatives are an unamendable core specified in the text, which India does not have, or nothing. This article does not resolve that dispute, because it is a genuine disagreement about institutional design rather than a factual error by one side. The doctrinal argument in full is set out in the article on the basic structure doctrine linked above, and a reader who wants to hold a defensible position on it should read the case for and against there rather than adopting either slogan.

The third is that every constitutional amendment requires the approval of the states. It does not, and this one is simply wrong rather than contestable. Ratification is required only for the provisions listed in the proviso, and the great majority of amendments have been enacted by Parliament alone. The confusion is understandable, since the ratification requirement is the most distinctive feature of the Indian procedure and the one most likely to be mentioned in a summary, but the consequence of the error is significant: it leads people to assume a federal safeguard exists over changes to fundamental rights, where none does.

How an amendment is drafted before anyone sees it

The parliamentary stage is public and the drafting stage is not, which is why accounts of amendment tend to begin at introduction and thereby miss where most of the decisions were taken.

A constitutional amendment ordinarily originates in the administrative ministry that owns the subject rather than in the Law Ministry. The ministry identifies a problem it cannot solve by ordinary legislation, usually because a court has held that the ordinary legislation would be unconstitutional or because the subject lies outside Union competence, and it prepares a proposal. The Legislative Department of the Ministry of Law and Justice drafts the text, and the Department of Legal Affairs advises on constitutional exposure. Where the proposal follows a commission or committee report, the report’s recommendations serve as the starting text and the gap between the recommendation and the draft is usually the most revealing document in the entire record. Cabinet approves the introduction. Only then does the bill become visible.

Two drafting decisions taken at this invisible stage govern everything afterwards. The first is the classification decision, meaning the government’s own view about whether the proviso is attracted, and that view is not always recorded or explained. The co-operative societies amendment is the standing example of a classification decision taken one way by the government and the other way by the courts a decade later, and there was no public argument about it at the time because the question was never squarely put. The second is the choice between amending the Constitution and legislating under existing powers, which is more often a political choice than a legal one: entrenching a policy in the Constitution makes it harder for a successor government to reverse, and that durability is frequently the actual objective rather than a side effect.

Reading an amendment properly therefore means reading three texts against each other: the recommendation that proposed it, the bill as introduced, and the Act as passed. The distance between the first and the second shows what the government wanted. The distance between the second and the third shows what it had to concede to get the numbers. That comparison is the analytical core of every legislative history article in this series and it applies with particular force to constitutional amendments, because the concessions are visible in constitutional text and are therefore permanent.

Ranking the provisions by how hard they are to change

A useful way to consolidate everything above is to ask which parts of the Constitution are genuinely difficult to alter, since the answer does not match intuition.

The easiest are the provisions the Constitution expressly left to ordinary legislation. The map of India, the existence of second chambers in the states, the number of judges on the Supreme Court, the salaries of members of Parliament, and the administration of scheduled and tribal areas can all be changed by a government with a working majority on an ordinary day. These are not trivial matters. Redrawing state boundaries has been among the most consequential exercises of legislative power in Indian history, and it has never once required a special majority or the consent of the state being divided.

The next tier is most of Article 368 territory: fundamental rights, directive principles, the composition and powers of Parliament, the emergency provisions, and the Preamble. Changing these requires the double majority in both Houses, which is a real constraint on a government without Upper House numbers and no constraint at all on one that has them and can hold its own members. Nothing federal stands in the way. The chapter on fundamental rights, which most readers would nominate as the most protected part of the document, sits in this tier, and it has been amended many times.

The third tier is the federal provisions in the proviso. These require everything the second tier requires plus the agreement of half the states, which converts the exercise from a parliamentary negotiation into a national one and lengthens the timeline by months at minimum.

The fourth tier is not a tier of provisions at all but of features, and this is the point at which the neat hierarchy breaks. The basic structure limitation does not attach to identified articles. It attaches to structural features that may be expressed across several provisions or in none in particular, and it can therefore render a technically simple amendment impossible while leaving a technically difficult one untouched. A change to a single word may be void if the word carries a structural feature, and a change to an entire Part may be valid if it does not.

That mismatch between the procedural hierarchy and the substantive one is the deepest structural fact about the amending power. The Constitution grades its own procedure by subject matter, and the courts grade the substantive limit by structural significance, and the two gradings do not line up. Every serious argument about constitutional amendment in India is, at bottom, an argument about that misalignment.

What the amending power gets right and where it is badly built

A pillar article owes the reader a verdict rather than a summary, and the verdict here is mixed in a specific and describable way.

The design gets three things right. The graded procedure is the most important. Requiring the same effort to redraw a state boundary as to alter the fundamental rights chapter would have produced either a paralysed federation or a devalued rights guarantee, and the Assembly avoided both. The federal proviso is the second: entrenching the Union-state bargain against unilateral parliamentary alteration, while setting the threshold at half the states rather than two-thirds or three-quarters, gives federalism a real protection without giving any small group of states a veto over national reform. And the entrenchment of Article 368 within its own proviso is a piece of drafting that has done more work than its length suggests, because it is the reason Parliament cannot quietly lower the bar for everything else.

Four defects are equally clear and are worth naming plainly rather than being smoothed over.

The first is the absence of any time limit on ratification. A bill can wait indefinitely for the resolutions of the last few states, and the government controls when it is presented to the President. Nothing prevents ratifications separated by years from being aggregated, which means a state legislature can bind its state on the strength of a resolution passed by a differently composed House under a different government. Most codified constitutions that use a ratification mechanism specify a window. This one does not, and the omission has never been tested because governments have not needed to test it.

The second is the silence on withdrawal. If a state may not withdraw a ratification, then a resolution passed in error or under a government subsequently voted out is irrevocable. If a state may withdraw, then the count is unstable until the moment of presentation and the government’s timing becomes strategically decisive. The Constitution answers neither way and the question is live.

The third is the vagueness of the trigger. The proviso applies where an amendment “seeks to make any change” in the listed provisions, and the co-operative societies litigation established that this is judged by substance. That is the right result, but it leaves governments drafting into uncertainty, since there is no way to know in advance how substantial a restriction of state legislative field will be held to constitute a change in the distribution of legislative powers. An amendment can be void for a procedural failure that was not identifiable as a procedural failure when it was drafted.

The fourth is the absence of any mandatory deliberative stage. There is no requirement of pre-legislative consultation, no requirement of committee referral, no requirement of a minimum period between introduction and passage, and no requirement that the states be consulted before an amendment affecting them is introduced, as distinct from ratifying it afterwards. A constitutional amendment can be introduced and passed in both Houses within days, and amendments have been. The safeguards in the Indian procedure are almost entirely arithmetical, and arithmetic is a poor substitute for deliberation when a government has the numbers.

A fifth point sits between a defect and a design choice, depending on the reader’s priors. The procedure gives the states a role in amending the federal provisions and no role whatever in amending anything else, including the chapter on fundamental rights, which every state administers and enforces. A change to the free speech guarantee or to the equality guarantee reshapes what every state government may do, and yet no state legislature is consulted at any stage. The Assembly’s reasoning was that rights belong to the citizen against the state and are therefore not a matter for bargaining between governments, which is a coherent position. The consequence is nonetheless that the most federal-looking safeguard in the amending procedure protects the division of powers and not the citizen, and readers who assume otherwise will misread both the proviso and the amendment record.

The overall verdict, then, is that the Indian amending procedure is well designed for the problem the Assembly was solving, which was how to permit a new state to remake its legal order without permitting the government of the day to remake the state. It is less well designed for the problem that emerged later, which is how to ensure that constitutional change is deliberated rather than merely counted. The judicial limitation has ended up performing the deliberative function by default, which is a substantial part of why it is contested: a doctrine that exists partly to compensate for a procedural gap will always look like an intrusion to those who think the procedure was sufficient.

Article 368 read against other amending formulas

A short comparative check is useful because the Indian procedure is often described as unusually flexible, and the description needs qualification.

Against the American formula, which requires two-thirds of both chambers of Congress and ratification by three-quarters of the states with no judicial limitation on the content of a valid amendment, the Indian procedure is more flexible on the federal axis, since half the states rather than three-quarters suffices and most amendments require no state involvement at all, and less flexible on the substantive axis, because the basic structure limitation has no American equivalent.

Against the Westminster position, where constitutional arrangements can be altered by an ordinary Act of a sovereign Parliament, the Indian procedure is substantially more rigid, and the comparison is worth making because the Indian Constitution inherits a great deal of parliamentary form from that tradition while rejecting its central constitutional premise. Parliamentary sovereignty is precisely what Article 368 read with the basic structure limitation denies.

Against constitutions that entrench certain provisions permanently, placing them beyond amendment by any procedure, India occupies a middle position that is easy to misdescribe. There is no eternity clause in the Indian text. Every provision is amendable by the procedure, and the limitation is not on which provisions may be touched but on what the cumulative effect of the change may be. That is a meaningfully different arrangement from an unamendable list, and it is the reason the basic structure limitation cannot be defeated by pointing out that the Constitution declares no provision unamendable.

The comparative lesson is that the flexibility of a constitution cannot be read off its amending clause. It is a function of the amending clause, the party system that determines whether the required majorities are assemblable, the federal structure that determines whether the states can be persuaded, and the judicial practice that determines what survives. India has a moderately demanding amending clause, a party system that has periodically delivered the required majorities, a federal structure that has proved negotiable, and a judiciary that has intervened rarely but decisively. That combination, not the text of Article 368 alone, explains the amendment record.

The errors that recur, and how to avoid making them

Because this subject is summarised more often than it is studied, a stable set of errors has propagated through general writing and into a good deal of teaching material. Each has a short correction and each correction is worth carrying.

Describing the special majority as two-thirds and stopping there is the most common. The omission of the total membership limb matters most precisely when it matters: in a thin House, where two-thirds of a small number of voters can be far below a majority of the seats. Any description of the procedure that does not contain both limbs is incomplete in a way that changes the answer.

Treating the simple majority route as an amendment under Article 368 with a lower threshold is the second. It is not a route within Article 368 at all, the Constitution says so expressly in several places, and the difference is not pedantic: it determines whether the ratification proviso can ever be engaged and whether the change counts in the amendment numbering at all.

Implying that a joint sitting can resolve a deadlock on an amendment bill is the third, and it is the error with the largest practical consequence, because it converts an absolute veto into a delaying power in the reader’s mind. Everything about how amendment negotiation works follows from the absence of that mechanism.

Assuming that ratification is required for major amendments and not for minor ones is the fourth. The proviso is drawn by subject matter, and the subject matter is federal. Importance is irrelevant to the classification.

Stating that an amendment is in force because it received assent is the fifth, and it is the error most likely to appear in otherwise careful writing. Commencement follows the terms of the amendment Act, which frequently defers operation to a notification, sometimes provision by provision.

Reporting bill numbers as amendment numbers is the sixth. The sequences diverge and the divergence widens over time.

Finally, describing the Constitution as either rigid or flexible without qualification is the error that contains all the others. The document is deliberately both, in different places, by design, and the whole purpose of learning the routes is to know which description applies to the change in front of you.

Amendment and interpretation as alternative routes to constitutional change

One further point belongs in any complete account of the amending power, because it explains why the amendment record understates the rate at which the Indian Constitution changes.

Constitutional text can change by amendment. Constitutional meaning can change by interpretation, and in India it has changed by interpretation at least as often and sometimes more consequentially. The content of the right to life has expanded enormously through judicial reading without a word of the provision being altered. The relationship between fundamental rights and directive principles has been reworked judicially. The scope of the equality guarantee has been rebuilt case by case. None of this appears in a count of amendments.

The two routes interact in ways that are central to the political history of the Republic. An interpretation the government dislikes can be answered by an amendment, and the early amendment record is largely a sequence of such answers. An amendment the courts consider excessive can be answered by a doctrine, which is what happened in 1973 and again in 1980. And an amendment that survives can be narrowed by interpretation afterwards, which is the quietest and most common outcome of all: a provision inserted with wide language and read down in application until it does considerably less than its text suggests.

For a reader trying to understand what the Constitution requires on any given question, this has a practical implication that is easy to state and easy to forget. The text of a provision, including a recently amended provision, is the beginning of the inquiry and not the end of it. The record around the provision, meaning what it replaced, what was said when it was introduced, what was conceded to get it passed, and what the courts have made of it since, is where the operative content sits. That proposition is the organising thesis of this series and the amending power is its clearest possible demonstration, because Article 368 is a short paragraph whose actual operation is governed almost entirely by material that is not in it.

What this means depending on who is reading

The amending power looks different from four positions and it is worth being explicit about each, because the same body of material serves them differently.

For an advocate, the operative content is the proviso and its consequences. The classification question is the one that generates litigation, and the co-operative societies decision established that it is decided on substance rather than on the face of the drafting. That opens a genuinely arguable line in any case involving a constitutional amendment that restricts what state legislatures may do without formally touching the Seventh Schedule, and it opens a second line on severability, since the court that decided the point divided on how much of an unratified Part could be saved. The commencement position is the other practical trap: a provision that has been enacted but not notified cannot be relied on, and the assent date is not the operative date. Nothing in a general article of this kind can substitute for advice on an actual matter, and the classification and limitation questions in constitutional litigation are unforgiving of assumptions, so the position stated here should be treated as the framework for a question rather than the answer to one.

For a competitive-examination aspirant, the retrievable structure is the three-route classifier and the four gates. The two facts most likely to distinguish a correct answer from a plausible one are the conjunctive nature of the special majority and the unavailability of a joint sitting, because both are routinely tested and both are routinely got wrong. The proviso list is worth knowing by subject matter rather than by article number, since questions are usually framed in terms of what an amendment does rather than which article it amends. The distinction between the simple majority route and Article 368 is the single highest-yield conceptual point in this topic, and the cluster’s dedicated study map for constitutional amendments works through how the material has been tested and where marks are lost. Building the classifier into a revision structure of your own is more useful than reading it once, and it is worth keeping your own statute and case notes privately on VaultBook so that the classifier, the proviso list and the four gates sit in one place you control rather than scattered across handouts. When the structure is in place, working through previous-year questions on the ReportMedic explorer shows how the same four facts are asked in different clothing, and a daily practice habit on ReportMedic is what converts recognition into recall.

For a legislator or a legislative staffer, the useful content is the arithmetic and the failure map. The crossover point between the two limbs of the special majority determines whether a full House helps or hurts. The Rajya Sabha veto determines the negotiation. The dissolution rules determine the calendar, and a federal amendment introduced late in a Lok Sabha is at serious risk regardless of its merits. And the classification decision taken at the drafting stage determines whether the government needs a state strategy, which takes months to build and cannot be improvised after passage.

For a historian or a journalist, the amendment record is a running index of what the state was trying to do and what it could not do by ordinary means. Every amendment is evidence that a government encountered a constitutional obstacle, and the sequence of obstacles is a more honest chronicle of the Republic’s priorities than the sequence of manifestos. The most revealing question to ask of any amendment is not what it changed but what it was answering.

Where to go next in this cluster

This article is the process spine for everything that follows. It deliberately stops short of the doctrinal argument, the individual amendments, and the examination treatment, each of which is owned elsewhere.

For the judicial limit, meaning what Kesavananda actually decided as distinct from what it is popularly said to have decided, which features have been recognised as basic and why the list has never been closed, read the article on the basic structure doctrine. For the beginning of the amendment record and the pattern of amending in reply to a judgment, read the article on the First Amendment and why it came so fast. For the only attempt to place the amending power beyond review, and for the surprisingly large portion of that attempt that remains in force, read the article on the Forty-second Amendment. For the ratification proviso worked through a live federal bargain, including what the states extracted in exchange for their resolutions, read the article on the One Hundred and First Amendment and the goods and services tax council. For the ordinary legislative process this article has deliberately not restated, read the article on how a bill becomes law.

The verdict to carry out of this one is short. The Indian Constitution is not easy to amend and not hard to amend. It is graded, and the grading is by subject matter rather than by significance, which is why the classification question comes before every other question. Get the route right and the rest of the analysis follows. Get it wrong and everything downstream, including the arithmetic, the timeline, and the exposure to challenge, will be wrong with it.

The state legislature’s side of ratification

Almost everything written about the ratification proviso is written from the perspective of Parliament, and the state legislature appears in it as a counter rather than as an actor. Looking at the exercise from the other end fills in a part of the procedure that is otherwise invisible.

A state legislature asked to ratify is not passing a law. It is passing a resolution, which is a different instrument with a different character: it requires no assent, it creates no rights, it cannot be amended into something else and still count, and it is not subject to the ordinary legislative process within the state. The motion is typically moved by the state’s chief minister or the minister holding the relevant portfolio, debated for a few hours, and carried. In states where the ruling party at the state level differs from the party in office at the Union, the debate is where the state’s grievances about the amendment are placed on record, often at length, and the resolution is sometimes carried anyway because the state’s substantive objection has already been dealt with in the negotiation that preceded parliamentary passage.

Three features of that exercise deserve attention. The first is that the resolution is on the bill as passed by Parliament, which means the state has no power to modify anything. It ratifies or it does not. All the bargaining therefore has to happen before parliamentary passage, which is why federal amendments involve extended consultation with state governments at the drafting stage in a way that ordinary amendments do not. The consultation is political rather than constitutional, since nothing in the Constitution requires it, but it is functionally necessary because a government that reaches the ratification stage without having secured its states has already lost.

The second is the position of the second chamber in states that have one. The proviso speaks of the legislature of the state, and the practice has been for the resolution to be moved and carried in the legislative assembly. This has not been litigated and is unlikely to be, but it is one of several places where the procedure rests on settled practice rather than on text.

The third is what happens to a state that refuses. Nothing happens. There is no consequence for a state that declines to ratify, no obligation to give reasons, and no effect on the amendment provided half the states have agreed. A state can object in the strongest terms, vote down the resolution, and then find itself bound by the amendment because the count was reached elsewhere. That is a design choice with a clear logic, since a unanimity requirement would make federal amendment impossible, and it is also the reason the ratification proviso is a protection for the states collectively rather than for any state individually. The distinction matters in any argument that treats the proviso as a guarantee to a particular state, which it is not.

The Ninth Schedule as a device the amending power created

One category of amendment deserves separate mention because it accounts for a large share of the early record and because it illustrates how the amending power can be used to build a mechanism rather than to change a rule.

The Ninth Schedule was created by the First Amendment as a list of enactments protected from challenge on the ground that they abridge fundamental rights. Once the list existed, protecting a further statute no longer required a fresh constitutional argument. It required only an amendment adding an entry, which is a short bill with no substantive content of its own beyond a schedule of names. A substantial number of amendments across the following decades consist of exactly that: entries added, often in batches, mostly land legislation at first and later a broader range of statutes.

Procedurally these are entirely ordinary special majority amendments requiring no state ratification, which is worth noticing given that most of the protected statutes were state legislation. A state law could be, and frequently was, placed beyond fundamental rights challenge by a Parliament in which that state had a modest voice, without the state’s legislature being asked anything. That is a real asymmetry in the design and it followed directly from the fact that the Ninth Schedule is not among the provisions listed in the ratification proviso.

The device also shows why counting amendments is a weak measure of constitutional change. An amendment adding twenty entries to a Schedule is one amendment. An amendment altering a single word in the fundamental rights chapter is also one amendment. They are not comparable events, and any analysis that treats the amendment count as an index of constitutional stability is treating them as though they were.

What later happened to the protection the Schedule confers, and the date that divides Ninth Schedule entries into two classes for the purposes of judicial review, belongs to the article on the basic structure doctrine and is not restated here. The procedural point is the one that belongs in this article: the amending power can be used to create a mechanism that reduces the cost of future amendments, and once such a mechanism exists it will be used at a volume that the original debate did not anticipate.

What to check when reading any amendment Act

A short working discipline follows from everything above, and it is worth stating as a sequence because it turns the analysis into a habit.

Begin with the short title and the year, and check the year against the assent rather than the introduction, because bills carry a different number and often a different year from the Act they become. Then find the commencement clause, which is usually the second section, and establish whether the amendment came into force on assent, on a stated date, or on a date to be notified, and whether different provisions were given different dates. This single step prevents the most common error in the field.

Then classify the route. Read the substantive provisions against the proviso list and ask whether anything touches the federal provisions, remembering that the test is substance and that an amendment restricting the state legislative field can engage the proviso without touching the Seventh Schedule. If the proviso is engaged, the parliamentary record should show ratification, and the absence of any reference to it is a signal worth following.

Then compare the Act with the bill as introduced. The parliamentary record will show the amendments moved, accepted and lost, and the difference between the introduced text and the enacted text is where the political story of the amendment sits. An amendment passed without a single change from introduction was either uncontested or driven through; an amendment with substantial changes was negotiated, and the negotiation is usually more revealing than the debate.

Then read the transitional and savings provisions, which are routinely ignored and which frequently determine what actually happens. An amendment that alters an institution will usually say what becomes of its existing members, pending proceedings, and prior actions, and those provisions are the operative ones for anyone affected in the transition.

Finally, check whether the amendment has been litigated, and if so whether it survived whole, survived in part, or fell. An amendment struck down remains in the printed text with a footnote, and a reader working from the text alone will attribute to the Constitution a provision that has no legal effect. The amending power is the one area of Indian constitutional law where reading the document without the record is close to guaranteed error.

The amending power when Parliament’s own term is in question

One procedural question sits underneath the most contested passage in the amendment record and is worth isolating from the politics that surrounds it. Does a proclamation of Emergency alter the amending power?

Textually, no. Article 368 contains no exception for a period of Emergency, requires the same majorities, imposes the same ratification proviso, and offers no relaxation. A proclamation suspends nothing in the amending procedure and confers no additional constituent authority. Anyone expecting to find an emergency amendment power in the Constitution will not find one.

What a proclamation does change is the composition of the body exercising the power, and it does so through a provision elsewhere. The life of the Lok Sabha may be extended by law during the operation of a proclamation of Emergency, for a period not exceeding one year at a time and not extending beyond a stated period after the proclamation ceases. That extension is effected by ordinary legislation and requires no special majority. The consequence, when it has been used, is that a House whose electoral mandate has expired continues to sit and continues to exercise constituent power, while the ordinary correction mechanism of a general election is deferred.

That is a structural observation rather than a partisan one, and it holds regardless of who is in office. The Indian amending procedure protects against a government that lacks numbers. It does not protect against a government that has numbers and has also lawfully postponed the event that would test them. The safeguards in Article 368 are arithmetical, the arithmetic is drawn from the composition of the two Houses, and nothing in the article addresses how that composition came to be or how long it has been in place. The eventual response to the episode in which this became live was not a change to Article 368 but a tightening of the conditions for declaring and continuing an Emergency, which is examined in the article on the 44th Amendment and how far the Emergency changes were undone.

The counting problem: what actually counts as an amendment

A last point of hygiene, because the question comes up constantly and the confusion it produces runs through a great deal of published material.

Not every change to the printed text of the Constitution is a constitutional amendment. A States Reorganisation Act rewrites the First Schedule and adjusts the Fourth, and it is not an amendment. A law creating a state legislative council alters the arrangements in Part VI, and it is not an amendment. Changes made under the Fifth and Sixth Schedules to the administration of scheduled and tribal areas alter the Schedules themselves, and they are not amendments. In each case the Constitution says so in terms.

Conversely, some enactments that are formally amendments do almost nothing to the working of the Constitution, such as those that add entries to a Schedule or extend a time-limited provision for a further period. They are counted, and each adds one to the running total.

The consequence is that the number of amendments understates the amount of constitutional text that has changed and overstates the amount of constitutional structure that has changed, at the same time and for different reasons. When someone cites the total to make a point about the Indian Constitution, the useful response is to ask which amendments they mean, because the answer will almost always be a small subset that could be discussed on the merits rather than counted.

Frequently Asked Questions

Q: What is the procedure to amend the Constitution of India?

An amendment bill is introduced in either House of Parliament, by a minister or a private member, without any requirement of prior presidential recommendation. It must be passed in each House separately by a majority of the total membership of that House and by not less than two-thirds of the members present and voting. If the amendment touches the federal provisions listed in the proviso to Article 368(2), it must then be ratified by resolutions of the legislatures of not less than half the states before it goes further. The bill is then presented to the President, whose assent is mandatory, and on assent the Constitution stands amended in accordance with the terms of the bill. Certain changes to constitutional text follow a separate route by ordinary majority and are expressly declared not to be amendments under Article 368.

Q: Which constitutional amendments require ratification by state legislatures?

Only those that seek to make a change in the federal provisions identified in the proviso to Article 368(2). These cover the election of the President, the extent of the executive power of the Union and of the states, the Supreme Court and the High Courts, the distribution of legislative powers between the Union and the states, any of the Lists in the Seventh Schedule, the representation of states in Parliament, and Article 368 itself. The threshold is the legislatures of not less than one-half of the states, acting by resolution, before the bill is presented for assent. Importance is irrelevant to the classification: amendments to the fundamental rights chapter require no state involvement, while a technical adjustment to a Seventh Schedule entry does. The trigger is judged by substance, so an amendment that substantially restricts what state legislatures may do can attract the proviso even without altering Schedule text.

Q: Can a private member introduce a constitutional amendment bill?

Yes. Article 368 requires only that the amendment be initiated by the introduction of a bill in either House, and it imposes no requirement that the bill be moved by a minister or recommended in advance by the President. Private members have introduced constitutional amendment bills throughout the life of the Republic on a wide range of subjects. What they have not done is get one enacted, because the special majority in both Houses effectively requires the governing party to supply its numbers to a proposal it did not draft. The practical function of a private member’s amendment bill is therefore agenda-setting rather than legislation: it places a proposal on the parliamentary record, occasionally obliges a minister to state the government’s position on the floor, and sometimes reappears years later inside a government bill.

Q: What happens if the two Houses disagree on a constitutional amendment bill?

The bill fails. There is no mechanism to resolve the disagreement. The joint sitting procedure that exists for ordinary legislation, where the President may summon both Houses together and the matter is decided by a majority of those present and voting across the combined membership, does not apply to constitutional amendment bills, because Article 368 requires passage in each House separately by the specified majorities and a joint sitting could not satisfy that requirement. This gives the Rajya Sabha an absolute veto rather than a delaying power, which is the most consequential and least discussed feature of the procedure. A government that carries the Lok Sabha overwhelmingly and falls short in the Rajya Sabha has simply lost, and its only options are to negotiate the text, wait for the Upper House composition to change, or abandon the proposal.

Q: Can the President refuse assent to a constitutional amendment bill?

No. The clause was amended in 1971 to provide that the bill shall be presented to the President, who shall give assent, which removed whatever discretion the earlier wording might have supported. The President cannot withhold assent and cannot return a constitutional amendment bill for reconsideration, both of which are available in relation to ordinary bills that are not money bills. The Constitution prescribes no period within which assent must be given, so a theoretical question about indefinite delay remains unanswered, but no President has attempted to test it and the practice has been prompt assent once a bill is presented. Where ratification by state legislatures is required, it must be completed before the bill reaches the President, not afterwards.

Q: Has any constitutional amendment been struck down for want of state ratification?

Yes. The Constitution (Ninety-seventh Amendment) Act, 2011 inserted a detailed code governing co-operative societies into the Constitution without seeking ratification by any state legislature, on the view that the proviso to Article 368(2) was not attracted. Co-operative societies are a State List subject, and in Union of India v Rajendra N Shah, decided on 20 July 2021, the Supreme Court held by majority that the inserted Part was void for want of ratification so far as it applied to societies operating within a single state, because a substantial restriction of the state legislative field is a change in the distribution of legislative powers even where the Seventh Schedule text is untouched. The majority severed the Part so that it survives for multi-state co-operative societies, which fall within Union competence, and one judge would have struck the whole Part down.

Q: Which parts of the Constitution can be amended by a simple majority outside Article 368?

Those the Constitution itself left to ordinary legislation. The admission and establishment of new states and the alteration of the areas, boundaries or names of existing states, together with the consequential changes to the First and Fourth Schedules, fall here, and the Constitution states expressly that such a law is not an amendment for the purposes of Article 368. So do the creation or abolition of a state legislative council, subject to a resolution of the state assembly, the administration of scheduled areas and tribal areas under the Fifth and Sixth Schedules, citizenship provisions, the quorum in each House, the salaries and allowances of members, and parliamentary procedure. Parliament exercises ordinary legislative power in these cases rather than constituent power, which is why the ratification proviso can never be engaged by them.

Q: How many amendments to the Constitution have been passed since 1950?

More than a hundred, which is a high figure for a codified constitution of that age and is regularly cited as evidence either of flexibility or of instability. The raw count is a poor guide to either, because it treats every amendment as equivalent. A substantial share of the record consists of extensions of reservation of seats in the legislatures, adjustments to the schedule of official languages, changes consequential on state reorganisation, and technical corrections, none of which was contested. The amendments that reshaped the constitutional order are a much smaller group, clustered in identifiable periods: land reform and the property guarantee in the early decades, the confrontation over the amending power between the mid-1960s and 1980, and federal and institutional entrenchment from the late 1980s onwards.

Q: Can the Preamble to the Constitution be amended?

Yes, and it has been. The Forty-second Amendment altered the Preamble in 1976, inserting two descriptive words and changing the phrase about the unity of the nation. The legal question underneath, whether the Preamble forms part of the Constitution for the purposes of the amending power, was settled affirmatively in the Kesavananda litigation, displacing an earlier view that treated it as introductory matter outside the operative text. Being part of the Constitution, it is amendable by the ordinary special majority route, since it is not among the provisions listed in the ratification proviso. An amendment to the Preamble would nonetheless face the same basic structure limitation as any other amendment, which is a separate question from amendability and is where most public argument on this subject becomes confused.

Q: Does a constitutional amendment bill lapse when the Lok Sabha is dissolved?

It follows the ordinary lapsing rules. A bill pending in the Lok Sabha lapses on dissolution, as does a bill passed by the Lok Sabha and pending in the Rajya Sabha. A bill pending in the Rajya Sabha which has not yet been passed by the Lok Sabha does not lapse, because the Upper House is a continuing body that is never dissolved. Prorogation does not cause a bill to lapse. This matters more for constitutional amendments than for ordinary legislation, because an amendment requiring state ratification may need many months of negotiation, so a proposal introduced late in the life of a Lok Sabha is racing the calendar. The first attempt at the goods and services tax amendment lapsed on dissolution and had to be reintroduced as a fresh bill in the next Parliament.

Q: How is the special majority calculated when members are absent or abstain?

The two limbs behave differently, which is the point of having both. The total membership limb is fixed by the number of seats the House comprises and is unaffected by vacancies, absences or abstentions, so it is the same number on a full day and an empty one. The present and voting limb is calculated only on members who actually cast a vote, so absentees are excluded and a member who is present but abstains is not counted either. An abstention therefore lowers the two-thirds threshold by shrinking the denominator while leaving the fixed limb untouched. For a government short of the total membership figure, abstentions are worthless; for a government short only of two-thirds, they are quietly useful.

Q: Can the Constitution be amended without the Rajya Sabha?

No. Article 368 requires the bill to be passed in each House by the double majority, and there is no route that bypasses the Upper House. Unlike ordinary legislation, where a deadlock can be broken by a joint sitting in which the larger House effectively prevails, and unlike money bills, where the Rajya Sabha can only recommend, a constitutional amendment stands or falls in both Houses independently. Because the Rajya Sabha is elected indirectly by state assemblies on staggered terms with roughly a third retiring every two years, its composition lags a general election result by several years. A party that wins a landslide acquires the Upper House gradually through state election victories, if at all, which makes constitutional amendment a multi-year project for any government that begins without those numbers.

Q: Is there a time limit for state legislatures to ratify a constitutional amendment?

The Constitution prescribes none, which is a genuine gap rather than a settled design choice. The proviso requires only that resolutions from the legislatures of not less than half the states be passed before the bill is presented to the President for assent, and it says nothing about how much time may elapse between the first ratification and the last, or between parliamentary passage and presentation. In principle a bill could wait indefinitely for the final resolutions, and ratifications separated by a long period could be aggregated, which would allow a state to be bound by a resolution passed under a differently composed legislature. The point has not been tested because governments seeking ratification have generally moved quickly and sought comfortable margins rather than exactly half.

Q: Can a constitutional amendment be made by ordinance or by referendum?

Neither. An ordinance promulgated when Parliament is not in session has the same force and effect as an Act of Parliament, and an Act of Parliament made in the exercise of ordinary legislative power cannot alter the Constitution. Constituent power is exercisable only by Parliament in the manner Article 368 lays down, which requires a bill passed by the specified majorities in both Houses. There is likewise no provision anywhere in the Constitution for a referendum on any subject, so a popular vote has no constitutional role in amendment. Introducing a referendum requirement would itself require an amendment to Article 368, and since that article is listed in its own proviso, such a change would need ratification by half the states.

Q: What is the difference between constituent power and ordinary legislative power?

Ordinary legislative power produces statutes, which operate within the Constitution and are void to the extent they conflict with it. Constituent power produces constitutional text and alters the framework itself. Article 368(1) states expressly that Parliament acts in exercise of its constituent power when it amends, language inserted in 1971 to establish that an amendment is not a law for the purposes of the provision that voids laws inconsistent with fundamental rights. The distinction determines which procedure applies, which majorities are needed, and whether the ordinary voiding provision reaches the result. It is also why changes the Constitution permits by ordinary law, such as the creation of a new state, are not amendments even though they change constitutional text.

Q: Can a single amendment Act use more than one route?

An amendment bill can contain provisions that would individually fall in different routes, and when it does the higher requirement governs the whole bill. A bill that alters a fundamental right and also changes an entry in one of the Seventh Schedule Lists is not treated as two bills with two thresholds. It needs the special majority throughout and it needs ratification because of the federal component, and if ratification is not obtained the component that required it is exposed. This is close to what happened with the co-operative societies amendment, where the parts adding a freedom and a directive principle raised no federal question at all but the inserted regulatory Part did, and the courts severed the amendment along that line rather than treating it as a single indivisible object.

Q: Does the Constitution place any provision permanently beyond amendment?

No. There is no eternity clause in the Indian Constitution, meaning no provision is declared unamendable by any procedure, and this distinguishes the document from several constitutions that place identified articles permanently out of reach. Every provision, including the fundamental rights chapter, the Preamble and Article 368 itself, can be amended by the procedure that applies to it. The limitation the courts have developed operates differently: it does not protect listed provisions but prevents amendments that damage or destroy structural features of the constitutional order, which may be expressed across several provisions or in none in particular. That is why an amendment altering a single word can be vulnerable while an amendment rewriting an entire Part can be valid, and why the substantive limit does not track the procedural hierarchy.