Communism failed because it ignored human nature. The question the current era presents to us – the question that underlies the crises represented by the words Brexit, Trump and gilets jaunes, but will also outlast them – is whether liberalism has the same problem. Communism could not handle humans’ individual and familial self-interest. Can liberalism handle their inherent need to be part of a group that defines itself against other groups?

Groups define themselves against other groups not only in the sense that they distinguish themselves on the basis of what they are not but also, unfortunately, in the sense that they compete with those other groups for status and resources. This is inevitable. Liberalism is, at bottom, the conviction that state coercion – also inevitable – should be neutral between fundamental aspects of the identity of the citizens that are subject to it. Coercion can be justified, if at all, only if it serves the equal freedom of all who are subject to it.

As a result, liberalism in the 21st century has to advance an idea that will undoubtedly meet with resistance. Since modern states are as bound to be ethnically pluralistic as they are to be religiously pluralistic, liberalism must advocate separation between nation and state, just as it earlier fought for separation of church and state. The social fact that nations and states overlap but do not coincide leads inexorably, for liberals, to the normative conclusion that no state should belong to a single people. The ability of past liberals to avoid this implication of their basic principles depended on historical circumstances that are now passing away.

If this is right, then it is not surprising that liberalism is facing resistance, or that the triumphalist narratives of globalization and democratization from the 1990s look hollow. The concept of democracy and sovereignty that we inherit is bound up with the nation-state. If nations and states are to be separated, who is the we that makes democratic decisions? How does the state retain the loyalty needed to fulfill its functions? How do the ethnicities that have identified with that state for centuries understand themselves when their countries become postnational?

It has taken liberalism a long time to get to this principle, and as a pragmatic movement comfortable with power it will naturally try to soft-pedal the implications. But liberalism now faces a global countermovement and needs to get its foundational commitments straight. And pragmatism increasingly pulls in the same way as principle: the coalitions that could put liberal parties in power in the West do not belong to a single “people,” and they will want the policies they vote for to reflect that.

Many liberals have proposed “civic nationalism” as a halfway house. But this will not work. If civic nationalism involves no real ideological commitments, it is too weak to count as nationalism. But if it is strong enough to have any real ideological weight, civic nationalism is no more compatible with liberalism than the ethnic version.

While there is no doubt that problems and struggles lie ahead, liberalism does have resources to address this problem. Liberalism is the ideology best equipped to deal with “intersectionality,” the principle that one has multiple identities and that the way each identity is experienced depends on the presence or absence of the others. Intersectionality is usually associated with a radical moralism that does not fit well with liberalism, but this is a contingent fact that can be changed. With a less individualistic and a more intersectional understanding of why states need to be limited and pluralistic, liberalism could be an appealing philosophy for younger people in the West and could regain enough vigour to put up a fight against its populist enemies.

Liberalism and Nationalism

People need to belong to groups bigger than themselves or their immediate families, but smaller than humanity as a whole. And those groups necessarily define themselves by the fact that they are not part of another group. This is a phenomenon familiar to anyone who engages in political speech online and, indeed, to anyone who went to high school. According to paleoanthropologists, it was true of our ancestors on the East African savannah. Everyone has particularistic loyalties to “their own” – a phrase characteristic of George Grant, English Canada’s leading critic of liberalism – just because it is their own.

This is a problem for certain traditional liberal theories that focus only on the rights of the individual and the need for a state to define and protect those rights. The essential goal of that form of liberalism is to figure out how to constrain the state from becoming so powerful that it threatens the individual, while ensuring that it is powerful enough to protect individuals from one another. The classical liberal solution was a state governed by the rule of law and representative democracy, appropriately constrained by guarantees of individual rights.

In the 20th century, most liberals recognized that negative rights needed to be supplemented by progressive taxation and social insurance. In the English-speaking world this recognition was notably expressed in the 1942 report of Sir William Beveridge, a British aristocratic liberal whose work was enthusiastically embraced by the democratic socialist movement.1 They also recognized that the state needed to play a role in regulating total demand to avoid periodic economic crises, as taught by John Maynard Keynes, another liberal toff who became a source of intellectual inspiration for labour politicians. But the key point about the whole picture is that it did not specifically refer to any groups other than the state as a whole. Individuals would react primarily to economic incentives. States would be insurance companies with navies.

To be sure, liberals always emphasized the importance of freedom of association and freedom of religion as ways of guaranteeing group loyalties defined in contrast to the state. The foundational struggle for liberalism was to detach the state from a particular church, but the coalition in favour of doing this rested fundamentally on the social community of minority churches. Liberals welcomed voluntary communities as a source of sense of meaning and loyalty to their “own.” The only price of membership in the liberal state was that these groups must not coerce members who seek to leave and must not threaten the state itself.

There may be doctrinaire cosmopolitan rationalists somewhere who are offended by any claim of a community less inclusive than humanity itself. These people are bound to be disappointed by humanity’s tribalism, just as a doctrinaire communist would be bound to be disappointed on realizing that real proletarians were never going to be the “new socialist man.” People differ in how groupish they are, a measure of personality that psychologists label “openness” and can quantify as one of the five basic dimensions of personality. There has indeed been an evolution in the “WEIRD world” – Western, Educated, Industrial, Rich Democracies – toward higher and higher levels of openness with each generation. But no one – not even an Esperanto-speaking world federalist – can exist without a tribe. Liberalism prides itself on being a pragmatic way of thinking that does not seek to coercively impose a utopian vision on people, but rather to give them institutional space to decide for themselves. It therefore has to learn to live with this fact about human beings.

The trouble begins with the question of what sources of group identity legitimately hold the state together. Groupishness, as a universal human phenomenon, is not on its own enough to explain nationalism, which is not. For most of human existence, the groups that commanded loyalty and defined themselves against others were small enough for everyone to know one another. The territorial state as part of an international system is a product of European modernity, along with wage labour, the world market and the colonial empire.

Territorial states gradually undid the overlapping secular and ecclesiastical jurisdictions of medieval Christendom and replaced them with a single sovereign authority defined against other, similar sovereign authorities. These absolutist states needed to channel universal human groupishness into identities that secured their own cohesion. Modern institutions of public education were developed to try to reeducate the inhabitants of France and England (and later Italy and Germany) to be citizens of a country, in priority to all other smaller or larger loyalties. As the transformation of the Renaissance Kingdom of England into the19th-century United Kingdom of Great Britain and Ireland suggests, this process could not occur without violence and exclusion of stubborn attachments within.

In the 19th century, liberalism and nationalism were assumed to be allies. The “self-determination” of peoples seemed to be consistent with the self-determination of individuals. The idea of liberal nationalism was that each people would get its own state, once the “artificial” borders of traditional multinational empires had been broken up. The high point of this vision was Woodrow Wilson’s Fourteen Points. While it inspired many people at the end of World War I, it was fundamentally compromised by the reality of the Versailles Treaty and the use of the language of self-determination by the Nazis in their designs on multiethnic Czechoslovakia in 1938.

The trouble is that peoples do not conveniently locate themselves exclusively within contiguous borders. As a result, a state for one people is necessarily a state defined against some of those who live within it. Moreover, since history does not end and powerful forces drive peoples to move across borders, or cause them to have different rates of demographic increase, the ethnic relationships within the territory of the state will constantly change.

This problem could be ignored as long as those outside the ethnos but within the state could be ignored. This was never an option for countries like Lebanon, Belgium or Canada where no ethnic group could really triumph, but it could work as a matter of realpolitik in countries with numerically smaller ethnic minorities.

However, one of the features of liberalism is that it encourages internal critique, as the limit of the circle of equal, autonomous persons is expanded on the demand of those left outside it. Enlightenment liberalism was simultaneously a project of white bourgeois males and one making claims based on the situation of all human beings. This contradiction can be the basis for presentist condemnations of the racism and sexism of the Enlightenment project, but its more important consequence was that it provided rhetorical space for the excluded to demand change in the ruling elite’s own terms. The demands of equal liberty made by bourgeois white men have been rejected by leftist intellectuals, but actual progressive social movements embraced these demands while insisting that the scope of equality and liberty be expanded. Once this happened for peoples whose existence does not correspond to an existing (or even possible) border, the liberal answer to empire can no longer be nation, but rather some messy multicultural federation – a federation being a democratized empire.

Canada, for example, originated as the federal union of British North America. The Victorian conception of Britishness was complicated, involving racial mythology about Anglo-Saxons, the political economy of free trade, the science of the industrial revolution and the redescription of the common law as an instrument of individual freedom. “Britishness” meant different things to George Brown and to George-Étienne Cartier. But no matter what its exact connotations, the idea that any part of the world, no matter how distant geographically from the original islands, could be made British was an unmistakably imperial idea.

However, already with 1867, it was necessary to separate state and nation to accommodate the reality of a Catholic, French population that could neither be given full authority over a particular territory nor denied a share of political power altogether. This need to accommodate was not a given. It contrasted with how the British Empire treated the Acadians conquered in Queen Anne’s war in the early 18th century and with the hopes for assimilation expressed by Lord Durham – an English radical – in his 1839 report. After Durham, however, it was clear to Baldwin and LaFontaine that Canada could only be democratic on a binational basis.

This accommodation was originally offered primarily to French Canadians and to a lesser extent English-speaking Catholics, primarily of Irish descent. But compromises were also made with the Métis with the Manitoba Act and with the Indigenous groups of the west, at the nadir of their strength, with the numbered treaties. Although these promises were disregarded by the Canadian state with the full flowering of settler colonialism, they were not forgotten by those to whom they were made.

Since the 1960s, partly in response to Quebec nationalism and partly in response to its own increased diversity, English Canada has largely abandoned any British identity in favour of a “multicultural” one. At that time, English Canada expressed a nationalism directed primarily at the United States of America, and this nationalism remained politically salient up until the free trade election of 1988. But since then, urban English Canada has identified too closely with “Blue” America to be really nationalist, while the conservative belt of rural Canada has been more fertile ground for a populist nationalism that is ethnic in a broad sense. This has caused a counterreaction in urban Canada, which has basically rendered the idea of a Canadian identity based in a peoplehood untenable.

Francophone Quebec’s initial response to secularization and modernization was a thinly disguised ethnic nationalism inspired by anticolonialism, and formulated now in terms of language rather than religion and descent. Left Quebec nationalism has obviously not disappeared, but it is no longer the beating heart of progressive Quebec. As in English Canada, it is issues of immigration and assimilation that have the most resonance with populist nationalism.

Much of Indigenous Canada has embraced an anticolonial nationalism of its own. Some have disclaimed any identification with the Canadian state at all. But for pragmatists, at least, the real objective is to be integrated into the Canadian federal structure with an alternative source of sovereignty to that of the federal and provincial governments, along with tacit or explicit acceptance that the sovereignty so claimed is one that is shared with the transethnic institutions of the Canadian state.

Canada’s situation in these respects is not simple and is the consequence of its own history. But it also echoes developments throughout the world, where the cause of liberalism and the cause of postnational states have become more closely identified. Moreover, the struggle over the nonidentification of nation and state has increasingly replaced the 20th-century struggle between labour and capital or over the amount of government redistribution as surely as that struggle displaced earlier ones about the place of the throne and the established church.

Brexit is the perfect example. Its proponents see themselves as protesting against a federal Europe displacing the sovereignty of the United Kingdom. But its main obstacle has been how it has disrupted quasi-federal arrangements within the United Kingdom itself, particularly in Northern Ireland, unfortunately the laboratory of the identity conflicts of modernity from the 17th century to the 21st.

To the extent that the separation of nation and state becomes a core liberal value, it will face a backlash, which will not disappear with better economic times. National identity fits well with basic human groupishness. It has been central to both personal identity and state formation in the West, and in the world influenced by the West, for centuries. It is therefore hard to imagine that declining ethnic majorities will abandon nationalism, and the pretense that it is nonethnic will become increasingly thin. But since the “people” as defined by the populists will never really be all the people in the state, and since those excluded will easily perceive this fact, the ethnic majoritarian coalition will inevitably give rise to a countercoalition.

Since liberalism is fundamentally defined by the idea that the state should not enforce one particularist conception of the good against dissenters, it cannot really be neutral in this conflict, which is going to define politics for the foreseeable future. Just as liberalism emerged as a pragmatic response to religious diversity, while often having to manage unwieldy coalitions of dissenters from the dominant religion, now it must do the same with ethnic identity. Contemporary liberalism’s demand must be that nation and state be separated. Its base consists of those who are threatened by uniting them. Principle and strategy leave no retreat: liberalism allows ethnic identity, but it must deny that identity state power.

The Mirage of Civic Nationalism

But is there a compromise between modern liberalism and nationalism that we can live with once ethnic nationalism is excluded? Is there a “civic nationalism” that is demanding enough to represent an alternative to the ethnic variety, while being consistent with liberal principles? A number of writers worried about the threat of nationalist populism to liberal institutions – including Yascha Mounk, Francis Fukuyama and John Judis – hope so.2 Mounk, Fukuyama and Judis are all liberals and can all see that if nationalism defines itself by claiming that membership in the state should be coincident with membership in the nation, illiberal results follow. But they intervene to ask the “left” to embrace a “civic nationalism,” arguing that without a thick sense of national identity, there will not be the will to put together projects like the social welfare state.

If “civic nationalism” means nothing more than that it is good if the citizenry identify with the state they live in as a common enterprise, and reasonable that they expect it to look out for their interests, then it is consistent with liberal principles. In this sense, though, the “civic nation” plays no greater role than the “civic province” or “civic municipality.” Public-spirited Torontonians or Manitobans expect their local or provincial governments to look out for their interests. A patriotism about a country that is similar to that felt for one’s city can certainly be a benign sentiment that no liberal would quarrel with. But patriotism is an emotion, while nationalism is an ideology. An ideology must define itself against something. So “civic nationalism” – if it is worthy of the name – must define the nation in a way that excludes some “civic” perspectives.

As Americans, Fukuyama and Judis want something like a commitment to the Declaration of Independence, the Constitution and an optimistic, entrepreneurial attitude to life as an identity substitute for blood and soil. The dilemma is that any ideological identity that is thick enough to fulfill the emotional needs met by nationalism will be as exclusionary as an ethnic identity – and even more in conflict with the liberal commitment to free debate of ideas. The phrase un-American has a nasty connotation for a reason. For all the flaws of old Europe, and for all the problems with its essentially ethnic understanding of national identity, at least the concept of an “un-Dutch” idea makes no sense.

Let us take Canadian examples of the problem with an ideological conception of national identity. In Lament For A Nation, George Grant claimed that Canadian identity depended on a less individualistic and more deferential approach to social life than prevailed in the United States.3 As a result, he wrote the Liberal tradition off as hostile to Canada – even though it had been the dominant tradition since Laurier and, in the 19th century, had led to the development of responsible government and had been part of the grand coalition leading to Confederation. Grant had to distance himself from the obviously individualistic strains in the Diefenbaker Conservatism that he was defending. Not surprisingly, since Canada has always been a pretty individualistic place, Grant had to conclude that Canadian identity was doomed before it could start.

Grant’s Lament foreshadowed numerous attempts to tie public polices about which there should be debate in democracies to national identity, about which there cannot be debate. Grant did this with federal Crown corporations, a theme that their CEOs have taken up ever since. Liberals and social democrats did it with the Canadian model of medicare and, after 1982, with a Charter and model of judicial review borrowed from the United States. Conservatives responded relatively harmlessly by tying national identity to peewee hockey and coffee-and-donut chains and less harmlessly to a more militaristic foreign policy. The liberal objection to all this is that it makes support for or opposition to certain contingent public policies matters of loyalty to the state.

It is not clear that there is actually a positive relationship between a strong sense of national identity and social welfare. Countries that have long struggled with a common national identity – like Canada and Belgium – do not seem to differ in any important way on this dimension from countries that have not, like France and the United States. If the welfare state is an efficient means of delivering social insurance – and it is – then it is not clear why it would not be enough for its citizens to recognize this. It is an empirical question, and the empirical evidence is not very strong that a specific ideological commitment is necessary for people to be public-spirited.

More fundamentally, though, civic nationalism faces the strategic and political problem of having no constituency. The resistance to Trump and Brexit, for example, comes primarily from the people who feel most excluded from their definitions of “American” or “British.” An opposing coalition must consist in people who have a wide variety of incompatible identity commitments. Negotiating such a coalition requires bracketing various commitments and promising them some space in the public policy that will result if the coalition succeeds. Liberalism is good at creating that kind of space. One possibility would be to define the common denominator of the antiethnic coalition as the “true” civic national identity. But this would just further enrage the majoritarian populists: they would not be “true” citizens of their own country! It is better to recognize the inherent asymmetry in the two contending coalitions.

The Intersectional Liberal Federalism of George-Étienne Cartier

We cannot put the question off any longer. If neither a single ethnic identity nor a single political identity for a state is compatible with liberalism, then how does liberalism learn to live with groupishness? Are we stuck with the pessimistic conclusion that the principles of John Stuart Mill and Benjamin Constant are for a species with a different evolutionary history from our own, possibly descended from solitary gibbons? As Edward O. Wilson, one of the world’s experts on ants, said of communism, “Great idea, wrong species.”

One response is that no one ever said it would be easy. Liberalism, like socialism, has a tendency to think of its success as guaranteed by history, so that when inevitability is put in doubt, the alternative seems to be despair. A more realistic approach would be to keep normative commitments separate from short-term success or disappointment.

Still, liberals do need a strategy. An alternative might start with the observation, banal on the cultural left, that identities are “instersectional.” This much-mocked word contains two useful and yet undeniable insights. The first is that every person is subject to multiple particularist loyalties and experiences: we are not just women/men or Canadians/Americans, but Canadian women/American women/Canadian men/American men – in exponentially more specific intersections of these sets. The second is that the experience of being part of the same group differs depending on the other groups to which one belongs: African-American women differ from African-American men not only in their gender identity but also in how they experience their racial identity. This example can be generalized indefinitely. Intersectionality is, in this sense, an undeniable fact.

And it is a problem for nationalism of any kind. That is because nationalism needs to elevate one identity cleavage to supreme importance while diminishing all the others. For a consistent nationalist, one must be an American or a Pole, but not a Polish-American. At minimum, such fractures are threatening to the national identity and to the idea of one people. From liberalism’s perspective, however, this is good news. Its enemies have a problem with the species they belong to as well, since, in fact, people do not spontaneously keep to a single identity.

But liberals have generally been suspicious of intersectionality. One problem is that those employing intersectional vocabulary tend to confuse oppression with virtue. They will often treat every identity distinction as a vertical one of oppressor and oppressed, and never a horizontal one of groups that must share a common space. Moreover, they will explicitly say that the oppressor can never judge – or even understand – the claims of the oppressed. If taken to the extreme, this would mean that differences could never be justly resolved or even effectively negotiated. Liberals have always differed with radicals in that they doubt that a politics in which the perspective of the “oppressor” can be ignored entirely would be either just as a moral matter or likely to succeed as a prudential matter.

But many liberals who object to intersectionality fail to recognize that an acknowledgement that everyone’s identity is complicated constitutes the best argument against radicalism. Very few people would fit into all the “oppressor” boxes, and even fewer would be “oppressed” in every respect. Even at the individual level, everyone has to find more or less principled compromises. An intersectional radical cannot pretend that there will be a single revolutionary subject, like Marx’s proletariat.

Another problem for liberals is the worry that a focus on intersectionality will lead to despair about the possibility of communication and collective action. If we cannot talk across identity categories, or if statements must simply be accepted, then a virtually infinite proliferation of such categories would create a hyperindividualized nightmare of noncommunication. On this point, it is precisely the liberal tradition, which has long been focused on the problems of common governance across divides of commitment to comprehensive worldviews, that has the resources to be useful to those concerned with the intersectional nature of identity.

Canadians should be more familiar than they are with George-Étienne Cartier’s 1866 speech in favour of Confederation, in which he called for a new “political nationality” with which neither the “national origin” nor the “religion of any individual” would compete. In Cartier’s vision, this political nationality would be shared by people of all parties and was not intended to replace ethnic or religious loyalties. In Cartier’s exposition on the new federal scheme, the political nationality of being a Canadian would serve those interests where religious, linguistic or ethnic identity was irrelevant. Cartier accepted that English-speaking Protestant Upper Canadians, French Catholics, Irishmen and Maritimers would all need to be represented in the councils of the political nation; were he alive today, he would no doubt modify his list to include women, Indigenous people and visible minorities.

Cartier’s concept of a political nationality should be distinguished from a civic nationalism dependent on allegiance to a substantive political ideology. Just as ethnic, linguistic and religious nations would meet in the institutions of the new political nation to hammer out their differences, so too would ideological groups. No doubt Cartier’s approach presupposed that anyone engaging in sedition against the state order would be suppressed. But it did not require any greater commitment than a willingness to work with the institutions as they existed.

One advantage the Confederation generation had over us today is that the socially and economically dominant English-speaking Protestants of Canada West were able to conceive of themselves not only as the true British North Americans but also as a section within British North America. As such, they advanced their interests through their representatives in a framework that implicitly accepted that others would also advance their interests. This did not prevent various identity panics on the part of this group, from the Riel rebellion through the Manitoba Schools controversy to the Conscription Crisis in the First World War. But through all of this, a framework remained in which Protestant Ontarians participated as one (loud) voice among many.

By contrast, neither left nor right is comfortable viewing the declining demographic “majorities” of the West today as one identity group among many – with legitimate interests, but also with an obligation to compromise those interests with the interests of others. For right populists, these groups just are “the people” and their identity demands are the demands of the nation as such. “Race” or “ethnicity” is something that only the Other has, which implies that the majority is raceless and without ethnicity. If challenged, the declining majority identity points to its acceptance of the principle of colour-blindness in law and its openness to the support of members of minorities willing to assimilate unreservedly to the majority.

The left sees through this and is understandably reluctant to acknowledge the legitimacy of a majoritarian identity politics. However, the left then goes on to insist that the majority just accept the moral untenability of its own identity as the corollary of accepting that its identity is just one among many. Instead of just making a principled demand for the separation of nation from state, the left in effect asks one people to cease to exist altogether.

The challenge is how to turn declining majorities into participants in multicultural compromise. Only on the racist and fascist far right is the contradiction resolved in favour of an explicit advocacy of “white” interests, but of course any use of multicultural language from this corner is just cover. This problem is insurmountable so long as the majority ethnicity defines itself as “white,” or as “French”, “Dutch” etc. The political entrepreneurs who seek to redefine majoritarian concerns in terms that speak to identity will probably continue to claim to speak for “the” people, as Trump and the Brexiteers do. From a liberal perspective, this is no better than an explicitly racist appeal, because it dissolves the universal into the particularities of one group.

While denying the concept of human nature, communism did speak to some perennial human characteristics: a longing for collective action, a dislike of hierarchies of rank and status. Its failure was its inability to integrate these aspects of human nature with others. Nationalist populism speaks to the need to be part of a group, and the need for that group to be “one,” but it suffers from the reality that our identities are multidimensional. On the left, this has been understood as “intersectionality,” but this insight has suffered from the left’s lack of attention to institutional realism. Liberals should give up on trying to renew their tradition by going back to the well of a single national identity and instead embrace the multiplicity they are best placed to reconcile with social order.


1 Sir William Beveridge, Social Insurance and Allied Services (Cmd. 6404), presented to the British Parliament in November 1942.

2 See my review of Mounk’s The People vs. Democracy: Why Our Freedom Is in Danger and How to Save It and Fukuyama’s Identity: The Demand for Dignity and the Politics of Resentment in Big Thinking the Democratic Recession, Inroads, Winter/Spring 2019, and John Richards’s review of Judis’s The Nationalist Revival: Trade, Immigration, and the Revolt against Globalization in Cosmopolitan Errors, Nationalist Response.

3 George Grant, Lament for a Nation: The Defeat of Canadian Nationalism (Montreal and Kingston: McGill-Queen’s University Press, 1965).