A constitutional amendment is a formally authorized change to a constitution. It may add a provision, remove existing language, replace a rule, or alter the legal effect of part of the constitutional text. Because a constitution establishes public institutions, distributes governmental powers, and often protects fundamental rights, its amendment differs in authority and frequently in procedure from ordinary legislation. The term can refer both to the process of constitutional change and to the resulting legal instrument.
Amendment procedures address a central problem of constitutionalism: how to preserve fundamental commitments while allowing institutions and rules to respond to changing circumstances. Constitutions commonly require special parliamentary majorities, approval by territorial units, a referendum, or several of these mechanisms together. Some also place particular principles beyond ordinary amendment. Other constitutional systems permit important constitutional rules to be changed through ordinary legislative procedures.
The legal significance of an amendment therefore depends on more than the words adopted. It also depends on who may propose it, which institutions must consent, whether substantive limits apply, and how courts interpret its relationship with the rest of the constitution. Comparative study distinguishes formal amendment from constitutional interpretation, changes in political practice, and the replacement of an existing constitutional order.
Constitutional authority and historical development
Amendment is usually understood as an exercise of legally constituted authority: the constitution itself authorizes specified institutions to change its provisions. This differs from constituent power, the authority associated with establishing a constitutional order in the first place. The distinction explains why an institution that can amend a constitution may nevertheless lack authority to abolish its essential foundations. It also explains why compliance with voting requirements does not necessarily resolve every question about the validity of a constitutional change.
The development of written amendment rules accompanied the spread of modern codified constitutions. Article V of the United States Constitution, drafted in 1787, provided a durable model combining national proposal with state ratification. The French Constitution of 1791 adopted a more temporally restrictive revision process, requiring successive legislatures to express support for change before a special assembly could act. These arrangements reflected different concerns: the American procedure sought to make correction possible without allowing ordinary national institutions to act alone, while the French procedure sought to stabilize a newly established constitutional settlement.
A classic distinction separates “rigid” from “flexible” constitutions. A rigid constitution requires a procedure more demanding than that used for ordinary statutes; a flexible constitution does not necessarily do so. The distinction concerns legal procedure rather than the actual frequency of change. A constitution requiring a two-thirds parliamentary majority may be amended frequently where one party or coalition regularly controls that majority. Conversely, a legally straightforward change may be politically difficult because of entrenched expectations or a requirement to secure agreement among several parties. Legal rigidity and practical resistance to change are therefore related but separate characteristics.
Amendments also differ in their textual form. In the United States, amendments are appended to the constitutional document, leaving the original wording visible even where later provisions supersede it. Many other jurisdictions incorporate amendments directly into a consolidated text. Neither method determines the legal rank of the change: a valid amendment normally has constitutional authority regardless of editorial presentation. The distinction does affect interpretation, because an appended amendment makes the chronological relationship between provisions particularly visible.
Terminology is not uniform. “Amendment,” “revision,” and “alteration” may be interchangeable in one jurisdiction but identify different procedures in another. Switzerland, for example, distinguishes partial from total revision, while Spain prescribes an aggravated procedure for total revision and certain especially protected changes. The name attached to a measure is therefore less important than its scope, legal basis, and prescribed adoption process.
Proposal, legislative approval, and ratification
A formal amendment procedure commonly separates initiation, deliberation, approval, and entry into force. Initiation determines who may place a proposal on the constitutional agenda. Approval determines the level and distribution of consent required. Ratification, where separately required, gives another institution or electorate a decisive role after initial adoption. Publication, certification, or promulgation then establishes the authoritative text and may determine when the amendment becomes effective. Not every constitution separates these stages, and similar terms can describe different legal functions.
Article V of the United States Constitution provides two proposal routes: approval by two-thirds of both houses of Congress, or a convention called by Congress on the application of the legislatures of two-thirds of the states. Ratification requires approval by legislatures or conventions in three-fourths of the states, with Congress selecting the mode. All 27 ratified amendments were proposed by Congress. The Twenty-first Amendment, ratified in 1933, is the only one ratified through state conventions rather than state legislatures. The president has no constitutional role in approving an Article V amendment, so the ordinary presidential veto does not apply.
France illustrates a different relationship between legislative and popular approval. Under Article 89 of the Constitution of 1958, an amendment must first be adopted in identical terms by both parliamentary chambers. It then ordinarily proceeds to a referendum. For a government-sponsored proposal, however, the president may instead submit it to Parliament meeting in Congress, where approval requires three-fifths of votes cast. The distinction between a government bill and a parliamentary initiative matters because this alternative route is not available for every proposal. Agreement between the chambers is also indispensable under Article 89; the National Assembly cannot override the Senate as it can in some ordinary legislative circumstances.
Voting denominators are an important technical feature. A requirement of two-thirds of members present is different from two-thirds of all members, and both differ from a majority of the entire membership combined with a supermajority of those present and voting. Abstention may therefore have different effects across jurisdictions. Requirements for repeated readings, intervals between votes, or an intervening election add further conditions beyond the headline majority. Spain’s aggravated procedure, for example, requires approval in principle by two-thirds of each chamber, dissolution of Parliament, approval by the newly elected chambers, and a referendum.
Time limits create another set of questions. Some amendment proposals expire unless ratified within a prescribed period; others remain pending without an express deadline. The United States Twenty-seventh Amendment was proposed in 1789 and ratified in 1992, illustrating the consequences of a proposal without a ratification deadline. Deadlines, rules governing withdrawal of consent, and the authority responsible for certifying completion can become legally significant when approval unfolds across many institutions over an extended period.
Federal consent and territorial representation
In a federal system, amendment affects not only the relationship between public authorities and individuals but also the constitutional agreement among territorial units. A national population majority may support a change that substantially reduces the autonomy of particular states or provinces. Federal amendment procedures therefore often combine population-based approval with territorial consent. The resulting rules protect constituent units, although they can also give less populous units influence disproportionate to their share of the national population.
Australia’s Constitution, in section 128, ordinarily requires an alteration to pass both houses of the Commonwealth Parliament by absolute majorities before submission to voters. It also provides a route for submission after repeated passage by one house under specified conditions. At referendum, the proposal generally needs a national majority of votes cast and majorities in a majority of states: at least four of the six. Votes in the territories count toward the national result but not toward the state-majority requirement. Certain alterations affecting a state additionally require approval by voters in that state. These rules distinguish national democratic approval from consent distributed across the federation.
Switzerland similarly requires constitutional changes to receive both a majority of the popular vote and a majority of cantonal votes. The cantonal vote is determined by the popular result within each canton; six cantons have half a cantonal vote each. A proposal can consequently win a national popular majority but fail because it does not secure the required territorial majority. Swiss constitutional design combines this federal safeguard with extensive opportunities for popular initiation, making territorial approval part of a broader system of direct democracy.
Canada’s Constitution Act, 1982 establishes several amendment formulas rather than a single universal rule. The general procedure requires resolutions of the Senate and House of Commons and at least seven provincial legislatures representing at least 50 percent of the population of all provinces. Other matters require unanimity, including changes concerning the office of the monarch, governor general, or provincial lieutenant governor, the composition of the Supreme Court of Canada, and the amendment formulas themselves. Changes applying to some but not all provinces can follow a procedure involving the federal Parliament and the affected provinces. Limited categories of change can be made by Parliament or a provincial legislature acting alone.
These arrangements make classification decisive. Before counting votes, institutions may need to determine which formula governs a proposed change. Canada’s Supreme Court held in its 2014 Senate Reform Reference that Parliament could not unilaterally introduce consultative Senate elections or fixed senatorial terms, and that abolition of the Senate required unanimous provincial consent. The case illustrates how an amendment’s constitutional effect, rather than its legislative label alone, can determine the necessary territorial participation.
Referendums and public participation
A constitutional referendum gives voters a direct decision on a proposed constitutional change, but its legal role varies considerably. It may be mandatory for every amendment, required only for specified subjects, or available if a sufficient number of legislators or citizens demand it. Some votes ratify a completed text; others authorize a drafting process or select between constitutional alternatives. A public consultation without binding legal effect is distinct from a referendum that forms a necessary stage of amendment.
Ireland requires constitutional amendments to be approved by referendum after passage of a bill through the Oireachtas. Approval depends on a majority of votes cast, without a turnout quorum. The Thirty-fourth Amendment, approved in 2015, authorized marriage between two persons without distinction as to sex. The Thirty-sixth Amendment, approved in 2018, replaced the previous constitutional provision concerning the right to life of the unborn with language permitting legislation on termination of pregnancy. These examples show how a constitutional vote may remove an existing restriction or authorize later legislation rather than supply a complete regulatory code itself.
Italy uses an optional confirmatory referendum under Article 138 of its Constitution. Constitutional legislation must receive two deliberations in each chamber, separated by at least three months, and obtain an absolute majority of each chamber’s membership on the second vote. If the second vote falls short of two-thirds in either chamber, specified parliamentary minorities, 500,000 electors, or five regional councils may request a referendum within three months of publication. If each chamber approves by at least two-thirds on the second vote, no such referendum may be requested. A referendum held under this procedure has no turnout quorum, unlike Italy’s abrogative referendum for ordinary legislation.
Denmark combines parliamentary approval, an intervening general election, renewed approval by the newly elected Parliament, and a referendum. Under section 88, an amendment must receive a majority of votes cast and support from at least 40 percent of the entire electorate. Japan’s Article 96 instead requires initiation by a concurring vote of at least two-thirds of all members of each parliamentary house, followed by ratification through a majority of votes cast at a designated referendum or election. Such differences make the legal meaning of abstention and the practical burden of approval highly jurisdiction-specific.
Public participation can also precede formal ratification. Ireland’s Constitutional Convention and Citizens’ Assembly considered proposed reforms before parliamentary action and referendums. These bodies illustrate deliberative democracy: selected participants receive information, hear competing arguments, and deliberate before making recommendations. Their recommendations do not themselves amend the constitution. More generally, hearings, petitions, published draft texts, and civic education can affect the formation of a proposal without changing the legally designated authority that must approve it.
Substantive limits and protected constitutional principles
Some constitutions restrict not only how amendments may be adopted but also what they may change. These substantive limits are often called eternity clauses or unamendable provisions. They seek to preserve specified foundations against alteration through ordinary amendment procedures. Such protection does not make a constitutional order historically indestructible; it establishes a legal limit within that order. A revolution or the creation of a new constitution raises a different question from a valid amendment under the existing constitution.
Germany’s Basic Law provides a prominent example. Article 79(2) requires amendments to secure two-thirds of the members of the Bundestag and two-thirds of the votes of the Bundesrat. Article 79(3) prohibits amendments affecting the division of the federation into Länder, their participation in principle in legislation, or the principles laid down in Articles 1 and 20. These include human dignity and core features of the democratic, social, federal constitutional state. The protection concerns foundational principles rather than an absolute prohibition on changing every word within those articles. Article 146 separately addresses the Basic Law’s replacement by a constitution freely adopted by the German people.
Other constitutions protect different institutional commitments. France’s Article 89 declares that the republican form of government may not be amended. Brazil’s Constitution of 1988 prevents deliberation on amendment proposals tending to abolish the federal form of the state; direct, secret, universal, and periodic voting; the separation of powers; or individual rights and guarantees. Portugal’s Article 288 contains an extensive list of material limits, including national independence, the republican form of government, political pluralism, and the separation and interdependence of sovereign bodies. The protected subjects reflect each constitutional settlement rather than a single universal model.
Temporal and circumstantial restrictions serve a different purpose. Brazil prohibits constitutional amendment during federal intervention, a state of defence, or a state of siege. France restricts initiation or continuation of amendment proceedings when the integrity of its territory is jeopardized. These rules aim to prevent fundamental change under conditions thought incompatible with free constitutional decision-making. A restriction may operate even where the proposed amendment’s substance would otherwise be permissible.
Substantive limits generate difficult interpretive questions. An amendment may leave a protected institution formally intact while greatly reducing its practical independence. Courts may therefore have to distinguish alteration from abolition and identify the essential content of a protected principle. A further issue is whether an unamendability clause can itself be repealed before the protected provision is changed. Constitutional systems do not answer this uniformly: outcomes depend on the text, the scope of amendment authority, and applicable judicial doctrine.
Judicial review of constitutional amendments
Judicial review of amendments can concern procedure, substance, or both. Procedural review asks whether the constitution’s requirements were satisfied: the correct voting threshold, competent institutions, mandatory notices, referendum conditions, or required territorial consent. Substantive review asks whether an otherwise properly adopted amendment exceeds the amendment power. The latter is especially consequential because it permits a court to invalidate a measure carrying the same formal label as the constitution itself.
India developed the best-known basic structure doctrine. In Kesavananda Bharati v. State of Kerala in 1973, a thirteen-judge bench of the Supreme Court held, by a narrow majority, that Parliament’s power under Article 368 did not extend to destroying the Constitution’s basic structure. The judgment did not provide a single exhaustive catalogue of protected features. Subsequent decisions have identified principles including constitutional supremacy, judicial review, secularism, and the democratic character of government. The doctrine differs from an express eternity clause because the substantive limit is derived judicially from the constitution’s structure and the nature of amendment authority.
The Supreme Court applied that reasoning in Minerva Mills v. Union of India in 1980. It invalidated provisions of the Forty-second Amendment that sought to exclude judicial review of constitutional amendments and declare Parliament’s amending power unlimited, and it rejected a broader displacement of specified fundamental rights. These decisions made limited amendment power itself part of India’s basic constitutional structure. They also established that adding text to the constitution does not necessarily make the text immune from constitutional scrutiny.
Colombia’s Constitutional Court developed a related but distinct doctrine of constitutional substitution. Under this approach, an authorized amendment may modify the constitution but cannot replace its defining structural features through a procedure intended only for amendment. In 2010, the court invalidated legislation calling a referendum that could have enabled a second consecutive presidential reelection, identifying both procedural defects and a substitution of basic constitutional arrangements. The doctrine focuses on the boundary between reform and replacement rather than simply asking whether a particular existing provision has been contradicted.
South Africa supplies a contrasting example of highly differentiated textual thresholds. Section 74 of its 1996 Constitution requires support from at least 75 percent of National Assembly members and at least six provinces in the National Council of Provinces to amend section 1, which states founding values, or section 74(1) itself. Other amendments face different requirements depending on their subject. Courts can review compliance with these rules and with constitutional duties of public participation. Across jurisdictions, however, judicial authority is not uniform: some courts recognize substantive amendment limits, some emphasize procedural review, and some face express restrictions on reviewing constitutional amendments.
Formal amendment, interpretation, and constitutional replacement
A constitution can change in practical operation without any alteration to its formal text. Courts reinterpret provisions, legislatures create new institutions within existing powers, and political actors develop practices governing how legal authority is exercised. These processes are often described as informal constitutional change. They must be distinguished from formal amendment because they depend on different sources of authority and may be reversible through different means.
In the United States, constitutional doctrine concerning racial segregation changed decisively between Plessy v. Ferguson in 1896 and Brown v. Board of Education in 1954, although the text of the Fourteenth Amendment did not change. Brown held racial segregation in public schools unconstitutional under equal protection. The development illustrates how interpretation can transform the operation of an existing constitutional guarantee. It does not mean that courts formally amend the text; the court presents its decision as an interpretation of law already in force. Disagreement over interpretive method concerns how such authority should be exercised.
The United Kingdom demonstrates why constitutional amendment cannot always be identified with revision of one superior codified document. Its constitution comprises statutes, common law, institutional arrangements, and constitutional conventions. Major statutes, including the Parliament Acts of 1911 and 1949 and the Human Rights Act 1998, have changed constitutional arrangements through parliamentary legislation. Under the orthodox doctrine of parliamentary sovereignty, Parliament can make or unmake any law, although modern constitutional practice also involves devolution settlements, political commitments, and judicial recognition of constitutional statutes. The absence of a universal special amendment procedure does not mean that constitutional change is politically unconstrained.
Formal amendment must also be distinguished from constitutional replacement. Replacement establishes a new constitutional text or order, sometimes through procedures expressly authorized by the old constitution and sometimes through a constituent process outside its ordinary amendment rules. The distinction is not reliably measured by length. A short amendment could transform the structure of government, while an extensive technical revision could preserve its central commitments. Constitutional systems that distinguish partial and total revision therefore require a judgment about institutional identity as well as textual scope.
The relationship between these forms of change is important. Repeated judicial interpretation can reduce pressure for textual amendment by accommodating new circumstances. Conversely, an amendment may overturn a judicial interpretation or place a disputed rule beyond ordinary legislation. The United States Eleventh Amendment followed Chisholm v. Georgia in 1793 and changed the constitutional treatment of certain suits against states. Formal amendment thus operates within a wider constitutional system in which courts, legislatures, political practices, and constituent processes provide distinct mechanisms of adaptation.
Rights, institutions, and public finance as subjects of amendment
Amendments address both constitutional principles and concrete institutional problems. Some establish enforceable rights; others redesign elections, redistribute jurisdiction, authorize taxation, or correct technical defects. Their effects depend on the relationship between constitutional wording and implementing institutions. A declaration of a right may immediately invalidate inconsistent laws, require later legislation, or do both. An institutional amendment may change a voting rule directly while leaving administrative details to ordinary statutes.
The United States Reconstruction Amendments illustrate several forms of constitutional intervention. The Thirteenth Amendment, ratified in 1865, abolished slavery and involuntary servitude except as punishment for crime following conviction. The Fourteenth Amendment, ratified in 1868, established rules of national and state citizenship and restricted state action through guarantees including due process and equal protection. The Fifteenth Amendment, ratified in 1870, prohibited denial or abridgment of voting rights on account of race, colour, or previous condition of servitude. Each also gave Congress enforcement authority, connecting constitutional guarantees to legislative power.
Changes to suffrage and representation have frequently required constitutional amendment because constitutions themselves define the electorate or the composition of public institutions. The United States Nineteenth Amendment, ratified in 1920, prohibited denial or abridgment of voting rights on account of sex. The Twenty-sixth Amendment, ratified in 1971, protected the voting rights of citizens aged eighteen or older against age-based denial. The Seventeenth Amendment, ratified in 1913, shifted election of senators from state legislatures to the voters of each state. These amendments altered different components of democratic participation: eligibility, protection against exclusion, and the method of selecting representatives.
Public finance is another major field. The United States Sixteenth Amendment, ratified in 1913, authorized federal income taxation without apportionment among the states and without regard to a census or enumeration, addressing constitutional constraints arising from the treatment of direct taxes. Germany’s 2009 constitutional reform introduced the debt-brake framework into the Basic Law, including a structural borrowing limit for the federation and restrictions on Länder borrowing, subject to specified qualifications and exceptions. Such provisions constitutionalize aspects of fiscal policy that might otherwise remain within ordinary legislative discretion.
An amendment’s formal adoption does not guarantee immediate or complete practical realization. Enforcement legislation, administrative capacity, litigation, funding, and public compliance can all affect implementation. Constitutional prohibitions on racial discrimination in voting in the United States preceded the Voting Rights Act of 1965 by nearly a century. The distinction between legal enactment and effective enforcement is therefore essential to assessing an amendment’s historical consequences. Constitutional authority establishes the rule’s status, while institutions and political conditions shape how that rule operates.
Amendment design, continuity, and implementation
Amendment rules distribute authority across institutions and across time. Requiring a large parliamentary majority can encourage agreement beyond a governing coalition. An intervening election allows voters to respond before a proposal receives final approval. Territorial ratification protects the constitutional position of constituent units. A referendum supplies direct popular authorization. These mechanisms are not interchangeable: a proposal may have broad electoral support but weak territorial support, or overwhelming parliamentary backing without majority approval among voters.
Procedural difficulty is therefore multidimensional. The number of decision points matters, but so do their independence and sequence. Two legislative chambers controlled by the same disciplined majority may impose less practical resistance than a legislature and a separately elected executive with competing preferences. A requirement for approval after an election may be a significant obstacle in one party system and a modest delay in another. Comparative assessment must examine institutional behaviour rather than infer practical rigidity from numerical thresholds alone.
The structure of the proposal also affects constitutional decision-making. A single-subject rule can limit the combination of unrelated changes and allow voters to express a clearer preference on each. A package amendment can instead facilitate a negotiated settlement by combining provisions that different groups regard as mutually dependent. Switzerland requires popular initiatives for partial constitutional revision to respect unity of subject matter, among other validity conditions. The underlying legal concern is that voters should not be forced to accept an unwanted constitutional change merely to secure a separate one with which it has no sufficient connection.
Transition rules are particularly important where amendments alter offices, terms, courts, or territorial powers. A new electoral system may apply at the next scheduled election rather than immediately. A change in judicial appointment procedures may leave incumbent judges in office. A redistribution of powers may preserve existing legislation until replacement laws are enacted. Without such provisions, an amendment can create uncertainty over pending proceedings, officeholders’ authority, and the continuing validity of earlier acts. An amendment’s commencement clause and transitional schedule can therefore be as consequential as its principal substantive provision.
The legal record must also identify precisely what became constitutional law. Questions can arise over discrepant language versions, correction of drafting errors, certification of referendum results, or the date on which the final necessary approval occurred. Repeal presents related issues: an amendment may remove a constitutional prohibition without automatically reviving every earlier statute that depended on it. The Twenty-first Amendment’s repeal of national Prohibition in the United States, for example, ended the Eighteenth Amendment’s constitutional regime while leaving a distinct framework of state regulatory authority. Implementation consequently requires attention to the amendment’s interaction with existing statutes, institutional powers, and rules governing the temporal effect of legal change.