The Future of Multiracial Democracy Majoritarianism Without Majorities

ABSTRACT

By the paradox of "majoritarianism without majorities," the author argues that ethnic majoritarianism, in the form of nationalist movements and the institutionalization of majoritarian privilege, has become a defining feature of twenty-first-century democracy even though ethnic majorities do not exist as self-aware political communities. These majorities are merely census labels imposed on a conglomeration of minorities. This essay argues that the key force generating this paradox is the formulation of democracy as a system of majority rule. This formulation provokes cycles of ethnic majoritarianism, and the periodic remaking of the political community by turning insiders into outsiders. The solution to the problem of majoritarian nationalism therefore lies in reconceptualizing democracy as a form of "minority" representation, and protecting the rights of outsiders in order to protect the rights of insiders.

Majoritarian nationalism has become a defining feature of twenty-first–century democracy. In the twentieth century, ethnic nationalism was mostly a minority phenomenon that involved regionally concentrated groups within larger political units seeking independence or autonomy. Today it is a majority phenomenon, with ethnonationalist leaders everywhere trying to reimagine the nation in the image of majorities, excluding ethnic minorities in the process.

Majoritarian nationalists claim that ethnic majorities are in danger of becoming a minority or being treated like one. In India, Hindu nationalists claim that differential fertility rates are reducing Hindus' population share, and that constitutional safeguards for minorities discriminate against the Hindu majority. In the United States, white-nativist nationalists fear that white Americans will become a minority due to immigration and high birth rates among nonwhites. White-nativist nationalism in the United Kingdom is fed by similar fears of a demographic decline.¹

Majoritarian nationalism is now seen as a primary cause of that other defining feature of the twenty-first century: democratic backsliding. As Pratap Bhanu Mehta has written, the specter of nationalism poses "a serious threat to liberalism—and to democracy itself."² Democracy requires the protection of ethnic minorities. But ethnic-majoritarian nationalism risks turning democracies into ethnocracies, and bringing authoritarianism and violence. As ethnic majoritarianism rises, therefore, the quality of a country's degree of democracy falls.³

The sprawling scholarly literature on how democracies should respond to diversity has generated many ideas for protecting minorities and immigrants, but little on how to respond to majority fears.⁴ There is good reason for this. Most scholars discount these fears as those of a "majority with a minority complex."⁵ Further, the structural foundations of majority privilege, in legal, economic, and educational institutions, usually remain robust despite demographic shifts. The most common scholarly response to the question of how democracies should respond to majoritarian nationalism, therefore, is to strengthen safeguards for minorities and immigrants, advocate for the renewal of shared civic identities, and identify the conditions under which electoral politics can prevent or contain majoritarianism.⁶

I argue, however, that such solutions miss their mark because scholars have misdiagnosed the problem of majoritarian nationalism for five key reasons. First, our understanding of majorities is flawed: They do not exist, if by the term "majority" we mean a self-conscious political community. Every majority that is now being mobilized in the name of nationalism is but a census category that assigns a common name to a conglomeration of minorities.⁷

Second, to say that majorities do not exist is not to say that majoritarianism does not exist. Majoritarianism is real, while majorities are not. Majoritarianism exists in the democratic world today in the form of both majoritarian nationalism and majoritarian dominance—that is, the institutional enshrinement of majority privilege in law and public policy, the economic and educational systems, and the census.

Third, we have also misunderstood the nature of the relationship between democracy and ethnic majoritarianism. Democracy is indeed a political system that requires the protection of ethnic minorities. But it is also the main force in the periodic construction of the ethnic majorities that threaten those minorities. In this sense, democracy periodically generates its own supposed antithesis.

The root of the problem lies in the traditional formulation of democracy as a system of majority rule. This formulation propels and legitimizes cycles of ethnic majoritarianism. This is a dynamic process. The frontier between ethnic majority and minority is a moving one: Some groups constantly face the danger, or promise, of being shifted from one side to the other through the normal workings of democratic politics. Moreover, this moving frontier interacts with the one dividing "native" and "immigrant." It is not only new arrivals clamoring for entry who are defined as "immigrants"; so too, sometimes, are minorities who have lived in a democratic country for years, even generations, and yet face expulsion by the government representing the ethnic majority.

Democracy and ethnocracy, therefore, are not antonyms; they go together. Democracy generates, sustains, and periodically reinvents ethnocracy. Thus the majoritarian nationalism we see in the world today is not evidence of democracy's backsliding. It is evidence of the cyclical working out of the basic contradiction at the heart of democracy itself.

Fourth, we have misunderstood the nature of the danger that majoritarian nationalism poses, which comes not from strength but from weakness. It is precisely because majorities are so hard to will into existence as a political community that majoritarian nationalism usually takes an incendiary and violent form. Further, at risk are not just minorities who lie outside the majority but also some of those who have been incorporated into the majority category. In thinking about the dangers of majoritarian nationalism, therefore, we must be attentive not only to the terms of exclusion but also to the terms of inclusion.

Finally, the dangers of majoritarianism did not begin and will not end with the life cycle of majoritarian nationalism. These dangers reside in the institutional structures that predate the political mobilization of majorities as self-conscious political communities and will live on if and after the nationalism subsides.

The solution to the problem of majoritarian nationalism therefore lies first in revising our understanding of the term democracy itself. We should see it not as a system of majority rule but as one of fearful, anxious minorities jostling for a place at the table. The trick is to grant them that place as minorities, without forcing them into the straitjacket of a majority category. If groups in a democracy can stake a claim to rights as minorities, then they will be less likely to seek refuge under a majoritarian umbrella or to fear being left out of it.

Further, the salvation of democracy lies not only in protecting the rights of those inside the political community, whether majority or minority, but in expanding and protecting the rights of those outside it. Protecting the rights of outsiders helps to ease insiders' fear of expulsion—and gives those expelled a way of returning.

Manufacturing Majorities

The obvious rejoinder to my claim that majorities do not exist is to point to the data recorded by population censuses. These data demonstrate the existence of a majority in each of the three democracies I focus on here: the United States, the United Kingdom, and India.

The United States, according to its 2020 census, has a white majority of 57.8, 61.6, or 71 percent, depending on how we count: 57.8 percent is the percentage of non-Hispanic Americans who consider themselves white alone; 61.6 percent is the percentage of all Americans, including Hispanics, who consider themselves white alone; and 71 percent is the percentage of all Americans, including Hispanics, who think of themselves as white alone or in combination with other races. The United Kingdom has a white majority of 83 percent, based on censuses administered in 2021 and 2022 in England and Wales, Northern Ireland, and Scotland. Finally, India has a Hindu majority of 80 percent according to its 2011 census, the most recent for which figures are available.⁸ But a close look at these census categories reveals that they are just labels for a mosaic of minorities that have never come together as a political whole.

Still, to call majorities mere census categories does not mean that such categories are trivial. Censuses provide ethnic-majoritarian movements with their basic raw material. In all three countries, these categories are the starting point for ethnic-majority nationalism to try to meld their heterogeneous members into a coherent and self-conscious political whole. They have not yet succeeded. There has been a tendency toward majoritarian dominance in all three countries since their inception. But this happened before the emergence of majorities as unified political communities. Majoritarian privilege became entrenched first; majorities then grew up under its shelter.

Consider America's white "majority." Included among those categorized as "white" by the U.S. census are distinct groups formed around attributes such as language, income, occupation, region, religion, or party identification, not all of which are not counted by the census. Many of these attributes have become identities that are passed down through families over generations. These group identities are often spoken of as "fragments" or "divisions" of America's white majority, presuming the prior existence of a whole that is subsequently divided into parts. But this "majority" is not a whole and never has been. "White" is but a capacious label that covers scores of minorities that have never actually functioned as a single self-aware political community.

Wait a minute, you might say. U.S. history has been defined by the white majority's efforts to subjugate nonwhite minorities—beginning with the U.S. Constitution's denial of full personhood to enslaved people (Article I, Section 2), followed by genocidal violence against Native Americans, a civil war over slavery, and after slavery's abolition, social and economic segregation and other forms of white-supremacist violence against African Americans and other minorities. Once citizenship and voting rights had been extended to Native Americans and African Americans in the twenty-first century, new forms of subjugation were invented through the medium of the criminal-justice system.⁹

This is the crux of the paradox of majoritarianism without majorities. Majoritarianism was built into the founding of American democracy not by a mobilized political majority but by the 55 delegates who spoke for it. The formalization of white privilege in the law preceded, rather than followed, the unification or popular mobilization of whites. It became the basis on which new members were subsequently inducted into the "white" category via the census.

Does focusing just on race and ignoring other dimensions of identity reveal the existence of a coherent majority? No. According to the 2020 U.S. census, there are 104 race-based groups included in the count of "whites," none of which is a majority and most of which are tiny in number.¹⁰ Even if we focus purely on those of European origin—the most common everyday definition of "white" in the United States (and more restrictive than that of the U.S. census)—it would be hard to say that a racial majority exists in the United States. The five largest groups citing European origin alone or in combination are of English (46.6 million), German (45 million), Irish (38.6 million), Italian (16.8 million), and Polish (8.6 million) descent.¹¹ All these groups are in a minority.

Further, whites of Hispanic origin are now more numerous than many of those with a European background. Almost ten-million Hispanics identified as only white in the 2020 census, and more than 25 million identified as white in combination with some other race. This makes those of Hispanic origin alone or in combination the fourth-largest group among whites in America.¹²

The heterogeneity of groups included in the category "white" is critical to interpreting the support base of former president Donald Trump. According to surveys, a majority of those classified as "white" (here meaning non-Hispanic whites who identify only as white) voted for Donald Trump in 2016 and 2020 (54 and 55 percent, respectively).¹³ These figures seem to suggest that the white "majority" politically coalesced around Trump. But did all white Trump voters actually vote as whites? Although he refused to denounce white supremacists, Trump also tailored his appeals to many minorities within the "white" category, sending differentiated messages to the white working classes, evangelicals, rural voters, veterans, and others.¹⁴

Consider now the "white" majority of the United Kingdom, which accounts for 83 percent of the population according to the 2021–22 censuses. This "white ethnic majority" also contains a collection of minority identities, including the regional identities of English, Welsh, Scottish, and Irish, some of which are expressed in resurgent regional nationalisms of their own; religious identities such as Catholic, Protestant, Muslim, Buddhist, and Jewish,¹⁵ which have been acquiring minority consciousness as the importance of religion in Britain has declined; and class identities, which reflect not merely income or education but distinct cultural worlds. The U.K. census procedure offers a clue about the heterogeneous nature of the groups included within the "white" category. The figure of 83 percent was produced by aggregating the data from multiple censuses administered separately in England, Wales, Scotland, and Northern Ireland, even though the relevant question was designed differently in each.

The emergence of a white census majority is very much a postcolonial phenomenon in the United Kingdom. It dates to the 1991 census, when the category of "white" was first introduced under the guise of "ethnicity." Before this, the term "white" referred to race rather than ethnicity, and while race played a role in shaping Britain's national identity, a racial majority did not.

An ideology of racial difference (in interaction with colonialism) helped to shape British national identity since at least the eighteenth century. By characterizing the populations of the metropole as white and European, and the populations of the colonies as nonwhite natives, this ideology justified the rule of one over the other. But the role of race in the eighteenth and nineteenth centuries was to separate the population of the metropole from the population of the colony, not to separate majority from minority within the metropole. In fact, the point of asserting racial difference from the colony was to claim racial sameness within the metropole.

In the United Kingdom, as in the United States, racial privilege came first and settled easily into Britain's democratic institutions. A racial majority followed later. Furthermore, this white "majority" has not obtained the allegiance of the majority of its supposed members, if the 2024 vote share for Nigel Farage's right-wing populist Reform UK party is any indication. Nonetheless, its 2024 performance exceeded that of any of its predecessors: The party obtained 14 percent of the national vote, surpassing the Liberal Democrats to become the third-largest political party in Britain.¹⁶

But the four-million voters who cast ballots for Reform UK represent only a minority among those categorized as "white." They are also an unstable minority, represented by at least three different parties in the last decade (Reform UK in 2024, the Brexit Party in 2019, and the UK Independence Party before that) whose fortunes have fluctuated wildly, rising to 13–14 percent in 2015 and 2024, and plunging to 2–3 percent in 2010, 2017, and 2019.

Consider, finally, India's 80 percent "Hindu majority"—another "artefact of categorization," created by the census, that brings together a mosaic of minorities on a scale that dwarfs the majorities of the United States and United Kingdom. India's census divides the country's population into six major religious categories: Hindu, Muslim, Christian, Sikh, Buddhist, and Jain. It allocates more than three-thousand groups professing distinct religious beliefs, practices, and names to one and only one of these categories.¹⁷ It does this according to census codes developed between 1871 and 1872, when the first census was administered in British India, and the present, without attempting to stipulate a common "essence" that unites all those classified as Hindu or otherwise.¹⁸ To this extent, India's Hindu majority and its non-Hindu religious minorities both represent census-created collectives, not communities.

The collection of groups with distinct religious beliefs and practices included in the "Hindu" majority can also be classified along other dimensions of identity, including several thousand castes, several hundred tribes, and several hundred languages and dialects, all of which also have distinct regional and class identities. This "majority" has historically not functioned as a community in Indian politics, not even during the partition of India, which represented the high point of religious polarization in the twentieth century.¹⁹ It is only since the rise of India's current ruling party, the Bharatiya Janata Party (BJP), that there has been a consolidation of the Hindu majority as a political force.

This is not to say that majoritarianism had not emerged in India before this. Although India's constitution did not describe India as a Hindu homeland, it instituted a wider set of safeguards and affirmative-action policies for Hindu caste and tribal groups than for religious minorities. In fact, eligibility for some of these programs required claimants to profess a Hindu identity.

Further, the long-dominant Indian National Congress always had a strong Hindu-majoritarian strain, which began to flourish unchecked in the 1980s. It was a Congress government that first enacted legislation opening the door to treating religion as the basis of citizenship²⁰ and began patronizing the pro-Hindu Ayodhya movement, which called for replacing a historical mosque in northern India with a temple. Further, India's non-Hindu minorities have long been the disproportionate targets of postcolonial state violence in the form of both police brutality and counterinsurgency.

The BJP has taken Hindu majoritarianism to new levels. But even now, the so-called Hindu majority remains a politically heterogeneous collection of minorities that has only partially coalesced into political community. The BJP has never obtained more than 38 percent of the vote in parliamentary elections. Even if all the party's votes came from Hindus alone, that would still not equal the support of a majority of Hindus.

If, as I have argued, majorities are constructed as a numerical category through census categorization, and their numbers change as census categories change, then they cannot be said to be in numerical decline. They are better described as being in the process of redefinition. Further, if these majorities are only now in the process of becoming political actors, then they cannot logically be said to be in decline as a political force either. They are better described as being in the process of creation.

Democracy and the Making of Majorities and Minorities

Most people understand the term democracy—which comes from Greek words demos, the people, and kratos, to rule—to mean government by the people and for the people. But what precisely does it mean for a government to be authorized by the people? Competitive elections are usually thought to be the key institution that represents popular authority. In every transition to democracy—whether from military or apartheid rule, monarchism, or colonialism—democracy arrives when nonelected rulers make way for elected representatives.

But we typically use the term "democracy" to mean more than elections: a form of government in which rulers are accountable to the people and constrained by law, in which there is freedom of dissent, in which minorities are protected, and in which conflicts are resolved peacefully. These meanings may not always be explicitly articulated, but they become apparent in the breach. "That's not democratic," we say in the face of arbitrary power, when dissent is punished, and, in particular, when ethnic minorities are harmed.

These commonsense notions of democracy are reflected in the various definitions of democracy which can ultimately be organized into two broad families, electoral and liberal. These definitions differ in how they define competitive elections and which additional attributes they require, if any, to call a system a democracy. But they agree on two points: competitive elections are necessary for a political system to be called a democracy no matter what else is required, and the outcome of those elections is to be decided by majority rule.

This is true even in countries with electoral systems that try to constrain majoritarianism. Whether under winner-take-all or proportional-representation (PR) systems, winners tend to be chosen according to the majority principle—which, despite its imperfections, is seen as fair and transparent and better than the alternatives. The only difference is that in winner-take-all systems, the winner gets everything, while in proportional systems, the losers get something too. A PR system in fact reinforces the majority-rule principle, because if all groups deserve to have a say in who governs them based on their population share, then majority groups, by definition, have a fair claim to more say than everyone else.

Because majorities are required to win elections, the population at large—and not just their delegates—must be organized into majority and minority. This introduces a fundamental political inequality in democratic systems that we have not properly recognized. The conventional understanding of democracy is that it is a system that abolishes political inequalities. But in fact, democracy should be understood as a system that replaces them.

The main form of political inequality that democracy abolishes is despotism, in which a king or pope or general or viceroy has full deciding power and the common citizen has little or none. But democracy replaces despotic forms of inequality with demographic ones. By "demographic inequalities," I mean a hierarchy based on numbers in the population, in which majorities count for more than minorities. If majoritarianism is unimaginable without the census, then democracy is unimaginable without majoritarianism in this demographic sense.

The inequality between majorities and minorities can help to create or perpetuate social and economic inequalities. No one would be surprised to learn, for instance, that those who belong to political majorities have better access to government services, entitlements, and educational or employment opportunities than do political minorities. But in a democracy, the distinction between majority and minority is a fundamental inequality in and of itself. It represents the unequal allocation of power, in which the majority rules and the minority complies. Thus it also produces the unequal allocation of agency, respect, and status. Minorities can be subjects but never rulers. They quite literally count less than majorities do. This inequality would exist even if disparities in economic and social position did not.

Democratic theory has not ignored this problem in the least. There is a large literature that addresses it in two ways—by arguing, first, that for a democratic system to survive, majorities and minorities should be fluid in nature, and, second, that where minorities are not fluid, they should be adequately protected.²¹ Liberal democracies, the argument goes, do better on both dimensions than electoral democracies.

But the presumption is that these majorities and minorities are generated by forces that are independent of, and often precede, the establishment of democracy, and that while democracy does better than other political systems in managing these inequalities, it may not be able to do away with them altogether. In particular, democracies that are unlucky enough to have ethnic majorities and minorities are not likely to do very well at all. Even so, these groups would likely fare even worse under autocracy. As Robert Dahl puts it:

The issue is whether in the long run a democratic process is likely to do less harm to the fundamental rights and interests of its citizens than any nondemocratic alternative. If only because democratic governments prevent abusive autocracies from ruling, they meet this requirement better than nondemocratic governments.²²

But in fact, democracy does not merely respond to preexisting ethnic majorities and minorities; it creates them. It also creates an ethnic boundary between those who belong to the people in whose name authority is exercised, and those who are deemed outsiders. Further, the formation of ethnic majorities and minorities, and of insiders and outsiders, is not a one-time process but an iterative and interactive one.

"The people cannot decide," wrote Ivor Jennings, "until somebody decides who are the people."²³ The implication is that the political community precedes and authorizes democratic government. But in fact, democracies continually make and remake the people (and thus the majorities and minorities) whom they claim to represent—generally along ethnic lines.

Imagine a population that exists before democracy is established. Individuals in this population can be described by attributes on many dimensions and sorted accordingly into multiple and overlapping categories of varying numerical strength.²⁴ But once democracy, or the idea of it, is introduced, it becomes imperative for individuals in a population to sort themselves, or be sorted by political entrepreneurs, into majority categories, because that is how power and status will be allocated.

Those who end up as minorities do so not by choice, but because their attributes do not permit majority status. Some of these minorities may be able to fashion themselves as local or regional majorities. But the unluckiest of all are those groups that do not have the numbers to count at any level, national or regional or local: They run the risk of total powerlessness.

In principle, majorities can form on the basis of many kinds of attributes. But when the stakes of membership in the majority are so high, and determine both status and access to resources, it follows that citizens and politicians will make every effort to sort themselves into majorities with a good chance to persist for some time. This makes them especially apt to sort themselves into categories defined by ethnicity or other "descent-based" attributes, which are more likely than others to endure.²⁵

This is why democracy tends toward ethnocracy: Ethnic or other descent-based majorities often form in anticipation or in the early years of democratic politics, but even when majorities are not descent-based at the outset, there is every incentive to introduce descent as a criterion for membership later, so as to stabilize majority status.

The Battle for the Census

It is through the census—widely considered a "scientific" instrument to measure the size and features of a country's population—that ethnic majorities and minorities are made and remade.

Constructivist scholarship in anthropology, sociology, history, and political science has challenged the view of the census as an objective scientific instrument, demonstrating that censuses can shape what they seek merely to count. As Kenneth Jones notes about the census in British India, "[It] existed not merely as a passive recorder of data but as a catalyst for change as it both described and altered its environment."²⁶ The insights of this scholarship, however, rarely make it into popular understandings of democracy, or into democratic theory.

The central insight from scholarship on the census that must be added to democratic theory is this: If democracy is rule of the people, then the census is the ground on which the people are forged. Adam Przeworski writes of democracy as a political system in which elections settle the conflicts that were once settled by war.²⁷ But democracy does not end civil wars—it transforms and conceals them. The new civil wars created by democracies begin on the new battlefield of the census. The census, in turn, becomes the basis on which groups can stake claim to the ownership of a democracy.

The battle for the census is usually invisible and bloodless: It is commanded by the government and the groups who control it; the armies are made up of bureaucrats and enumerators, and the weapons are categories. But this battle can be as dehumanizing as armed warfare. Census categories can be a form of ethnic cleansing, capable of accomplishing what one author has called "genocide" by census redefinition.²⁸ They can also create new ethnic groups. Census warfare in turn feeds electoral warfare, which can indeed become a bloody affair.

In the United States, for example, the white racial majority as counted by the census today occurred through what historian Nell Irvin Painter calls the successive "enlargements" of whiteness in America: In 1790, the term "white" referred only to propertied Anglo-Saxon Protestant males. This was the original "core" category in the history of whiteness in America.²⁹ The first enlargement, from 1790 to the 1850s, dismantled property qualifications, expanding the category of "white" to include most white Anglo-Saxon Protestant men. The second enlargement, from the 1850s to the early twentieth century, incorporated Irish Catholic Celts and Germans. The third enlargement, in the aftermath of World War II, inducted Central European immigrants and Mexican Americans. Painter argues that a fourth "great enlargement" is taking place in the twenty-first century, with wealth replacing color as a marker of whiteness. But I would contend that we are still in the midst of the third enlargement—incorporating those of Hispanic backgrounds into the "white" category on a much larger scale.

The U.S. census "majority" is a product of these enlargements. Had the census been designed differently, it would have produced a different, and equally true, snapshot of the American population. The census could have, for instance, just as easily described the population as being made up not of a single dominant white race, but of many white races. This might have produced a snapshot of a population with many dispersed groups (like Tanzania's) rather than one dominated by a white majority. The census might also have employed a definition of race that emphasized admixture: That would have produced a snapshot of a racially blended population, rather than one sharply divided by a color line.³⁰

The fact that the census constructed the United States as a white "majority" nation is the result not of objective scientific processes, but rather of a political process in which "winning the census" served majoritarian democracy. As the statistical directives guiding the U.S. census now note, there are no objective criteria that justify counting race as singular rather than plural, or that justify counting races at all: "The racial categories included in the census questionnaire generally reflect a social definition of race recognized in this country and not an attempt to define race biologically, anthropologically, or genetically."³¹

Consider now the case of the United Kingdom, whose census counts whites using the term "ethnic origin" rather than "race" and, like the U.S. census, does not claim scientific validity or any kind of objective validity for its question on "ethnic origin."³² Yet the design of the ethnicity question in the census for England and Wales fuses whiteness with an English or Welsh ethnicity: A respondent who claims English or Welsh ethnicity must claim a white identity; and those who claim to be nonwhite cannot, by design, claim English or Welsh ethnicity.³³ It is the design of the census, therefore, rather than objective attributes, that associates England and Wales with a white ethnic majority. (The questions for Scotland and Northern Ireland are designed differently.)

Consider, finally, the case of India, where data from the very first census, conducted under colonial rule in 1871–72, showed a "Hindu majority." Yet the British census authorities never managed to come up with objective criteria for the label "Hindu" or, therefore, for the Hindu majority. They considered classifying the Indian population based on thousands of local groupings variously called castes, sects, tribes, and races, without listing them under the umbrella category of "Hindu religion." Doing so would have produced a picture of the Indian population that resembled a jigsaw puzzle of incommensurable minorities rather than one with a dominant majority defined by religion. Census officials also considered introducing a separate category for "debatable" Hindus, as distinct from "genuine" Hindus, which, apart from offending the potential "debatables," would likely have produced a snapshot of a "multipolar" rather than "majority dominant" population.³⁴

But the census was implemented in India as the electoral principle was being introduced. Representatives of the population were well aware of the importance of winning the census. Leaders who self-identified as Hindu resisted categorizations that would put their "community" at a numerical disadvantage. The census of independent India settled the question, producing an unambiguous Hindu majority and creating a linked hierarchy wherein certain sects, castes, and tribes could be allocated only to some religions but not to others.

This linked structure, together with government affirmative-action policies, locks India's Scheduled Castes, a collection of "lower" caste groups that make up 16 percent of the population, into the Hindu "majority." Immediately after independence, individuals could only be counted as Scheduled Castes if they also claimed membership in the Hindu category. And if they were not classified as Scheduled Castes, they would not be eligible for affirmative-action benefits. Eventually, members of the Sikh or Buddhist categories were permitted to claim membership in the Scheduled Castes as well. But to this day, Muslims and Christians cannot profess membership in the Scheduled Castes or claim the affirmative-action benefits.

To a lesser extent, government policy also locks "Scheduled Tribes," who constitute 8 percent of the Indian population, into the Hindu majority. Scheduled Tribes can claim any of the six major religions counted by the Indian census, or be included in the residual category of "other." But, despite efforts on the part of some groups, they cannot be counted as a separate religious group.³⁵ In practice, most members of the Scheduled Tribes are classified as Hindu.

In each of these cases, the line between minority and majority as drawn by the census is closely related to the line between insider and outsider. This line, too, is adjudicated by the census or some other form of population counting, in conjunction with citizenship policies. Census majorities might attempt to swell their own ranks by permitting entry selectively to some immigrants, as happened with the enlargements of the category "white" in the United States in the nineteenth and twentieth centuries, or to deplete the ranks of minorities by denying or stripping them of citizenship rights—for example, the U.S. Constitution's "three-fifths clause," and nineteenth- and twentieth-century government efforts to deny citizenship to Native Americans and Asian immigrants; or the United Kingdom's passage of laws, beginning in the 1960s, that suddenly made nonwhites whose family had been in Britain for generations aliens or second-class citizens, exemplified most recently by the Windrush Scandal, in which Caribbean migrants with the legal right to stay in the country were mistreated or deported; or India's passage of the Citizenship Amendment Act (2019), which discriminates against Muslim migrants on the path to citizenship, and in the process also renders precarious the rights of many citizens who cannot prove that they are not migrants.

Reformulating Democracy as Minority Rule

The argument so far suggests that we have misunderstood the danger posed by majoritarian nationalism. That danger, we are told, lies in majoritarian nationalist movements and those who lead them. The implication is that if these movements and their leaders were defeated at the polls, the danger would pass or at least be contained. Further, this danger is believed to be a function of the size and strength of the majority. Discontented minorities can pose localized threats. They are not strong enough to do more. But a discontented majority can destroy democracy itself by taking over the state and bullying minorities into submission. This is why majority rule is so often associated with terms such as "tyranny," "domination," and "steamroller politics."

The danger posed by majoritarian nationalisms, however, comes not from their strength but from their weakness. It is precisely because they represent a collection of minorities, and not a coherent whole, that they are hard to build and sustain. No one is more aware of this than majoritarian leaders. This is why they are so quick to demonize others and why majoritarian nationalism is so often associated with violence. Violence acts as a glue that keeps otherwise disparate minorities together by giving them a common enemy to unite against.

Consider the example of the racist violence that swept across British cities in August 2024. On July 29, three young girls were stabbed to death by a seventeen-year-old assailant. Shortly after, far-right activists mobilized in protest, spreading lies on social media, calling the assailant a Muslim asylum seeker (he was neither Muslim, nor an asylum seeker). Almost three weeks of anti-immigrant and anti-Muslim violence followed, with more than a thousand people arrested.³⁶ This violence should not be understood as a spontaneous uprising by members of a growing white "majority," but as the result of an attempt by majoritarian leaders to accelerate the construction of a majority that was not growing fast enough on its own by stoking the fear of outsiders.

It is also because of its weakness that majoritarian nationalism is so often associated with authoritarianism. Even though they stand at the helm of a majority, majoritarian leaders cannot trust mere numbers to keep them in power: They routinely turn to authoritarian tools to do that, including centralization of both party and government, and suppression of dissent and the media.

The solution to the problem of majoritarian nationalism certainly lies in stronger protections for minorities and immigrants who become the targets of mobilized majorities. But it also requires paying attention to the internal diversity of the majority itself, and creating the basis for undoing the link between internal minorities and their majority label. If the majority is a collection of minorities, then every majoritarian nationalism includes as well as excludes, and the terms of inclusion can sometimes be as oppressive as the terms of exclusion. Take India's Scheduled Castes and Scheduled Tribes, for example. Together, they constitute about a fourth of India's population and almost a third of the Hindu majority. As noted above, most members of these categories are "locked" into membership in the Hindu majority by the census. Those mobilized by the BJP, furthermore, are inducted on terms that encourage conformity with upper-caste symbols and practices—and the rise of Hindu nationalism has also been accompanied by acts of brutal violence against sections of these groups who do not conform.

If we understand majorities as undifferentiated actors, we cannot analyze the ways in which majorities induce or force the compliance of those who belong to it. Further, when we as scholars dismiss the fears of the majority writ large as groundless, we also dismiss the very real grounds for fear that some minorities within the majority may have: Doing otherwise can, in fact, smack of complicity in ethnic majoritarianism. Conceptualizing ethnic majorities as collections of minorities, by contrast, makes it possible to respond to the concerns of some of these internal minorities without being coopted into a majoritarian agenda.

Ultimately, however, majoritarian nationalism and its leaders are not the root problem. Their defeat, therefore, cannot be a lasting solution. Underneath nationalism itself lies the legal, economic, social, and political infrastructure of majoritarianism. This infrastructure preceded the creation of mobilized majorities and will likely outlast it. Undermining majoritarian nationalism therefore requires not merely the defeat of majoritarian leaders but the dismantling of that infrastructure. That infrastructure, in turn, stands on the majoritarian foundation of democracy itself. Unless this foundation is rebuilt, any measures to defeat majoritarian nationalism or dismantle majoritarian laws will be mere band-aids.

The most fundamental solution to the problem of majoritarianism, therefore, requires two things: first, reformulating our understanding of democracy itself to render it a minoritarian system—that is, one that allows all groups a place at the table as minorities rather than having to attach themselves to a demographic majority category; and second, according enforceable rights to those who do not belong to the political community. It is only with a robust set of rights for outsiders that the rights of insiders can be protected.