Introduction

In many regards, the Coalition Avenir Québec (CAQ) victory in the Québec 2018 provincial election is exceptional. We need to go back to 1966, about half a century ago, to find an election where Québec voters elected a party that is neither the Libéral Party of Québec (PLQ: Parti Libéral du Québec) nor the Parti Québécois (PQ). Only time will tell whether or not 2018 is an outlier in the PLQ/PQ two-party system or simply the starting point of a new partisan system. In any case, it is important to understand what drove Québec voters’ decision to support a party that never received such a mandate before. There is a clear common ground with previous elections: Québec nationalism remains an important and latent issue which is central to explain voters’ choices even though the issue of Québec’s independence from Canada was under the radar during the 2018 election. As our paper will show, on the top of support for Québec independence, nationalism also includes issues such as promoting the French language or public policies on immigration and remains a crucial feature of Québec electoral politics (Bélanger et al. 2018; Blais and Nadeau 1992; Pinard and Hamilton 1986, 1994).

This can be surprising at first sight considering that the usual threat of a referendum on Québec’s independence was neutralized during the 2018 election, as the Parti Québécois’ leader (Jean-François Lisée) publicly pledged to not hold a referendum in a first mandate. This context heavily clashes with the previous 2014 election where the former PQ leader (Pauline Marois) did not clearly discard the possibility of holding a referendum and where Pierre Karl Péladeau, a wealthy businessman owning major Québec media (newspapers, radio, TV, etc.), announced in a spectacular way that he was joining the PQ in order to separate Québec from Canada.

However, even with no threat of a referendum, other aspects related to nationalism still played an important role in explaining voting behaviour. All parties, but especially the Coalition Avenir Québec, who led the pre-election polls forecasting, presented a nationalist agenda in one way or another. Among other things, the CAQ claimed that Québec must reduce its immigration level in order to better integrate newcomers and advocated for stronger measures to protect the French language in Québec. Moreover, they pledged to integrate a “values test” for those who want to apply for permanent residency and proposed the secularization of public institutions, preventing public-sector employees in a position of authority to wear religious symbols in the workplace (e.g. judges, police officers and teachers). These are examples of pledges that are all related to collective values, the way Quebeckers conceive their nation and how to implement public policies allowing newcomers to become part of the Québec nation in a way that is consistent with their conception of their nation.Footnote 1

Given this exceptional occultation of the issue of Québec’s independence via a referendum, this raises the question of how to theorize Quebeckers’ nationalism and its influence on voters’ choice. The theoretical questions of how to conceptualize nationalism and how to define the concept of “nation” have raised important debates in the field of political theory. Below, we describe three different schools of thought on these questions. However, these different theoretical approaches tend to rely on more abstract considerations to support their respective conceptions of the nation. Empirical work in this area is much less common, and the perspective of the citizens themselves is often left aside.

We build on these theoretical debates and on our own empirical research to propose a novel approach to ethnic nationalism based on citizens’ view of what it takes to be part of the nation, combined with in-group preferences, and argue that these characteristics are central to improve our understanding of nationalism. Accordingly, we contend that a liberal nationalist approach to nationalism is better suited to explain nationalism in Québec and we raise important challenges for the two other theoretical families: cosmopolitanism and communitarianism. Thus, our research has both practical implications, in that it deepens our understanding of citizens’ vote choice calculus during the 2018 election, and important normative implications as we show that nationalism has to be considered as a valued and politically influential part of Québec’s identity, but can also remain inclusive to a certain degree.

In the next sections, we first review the three major theoretical approaches to the nation and nationalism. We then link that literature to the case of Québec and detail our empirical strategy before digging into our public opinion results. Our findings show that some factors (such as being able to speak French) are perceived to be very important to be part of the nation, while others are seen as much less important (such as being Catholic). Second, we examine what factors are associated with greater approval of particular views of the nation. Surprisingly, most gaps between groups appear to be small, with some exceptions such as education levels or left–right ideology. Third, we analyse whether the conception of the nation had an impact on citizens’ vote choice. Our results are quite neat: citizens who are closer to an ethnic conception of the nation are much more likely to support the Coalition Avenir Québec and somewhat less likely to vote for the Libéral Party of Québec or Québec Solidaire. Surprisingly, it does not affect the citizen’s proclivity to support the Parti Québécois.

Cosmopolitan and nationalist approaches to the nation

From the political philosophy literature, we can roughly distinguish between three major approaches on the question of the importance and significance of nations. These three perspectives are (1) cosmopolitanism, which reject its non-instrumental importance, (2) liberal nationalism, which try to reconcile the importance of belonging to a particular nation with individual freedom and autonomy, and (3) communitarianism which puts forward the intrinsic value of nations, sometimes at the expense of individual freedoms and liberties. All three approaches have been proposed and endorsed by different influential thinkers and, we argue, are relevant to understand how one should think of the political importance of nations and nationalism. Below, we review the main features of each school of thought and illustrate how they have been deployed to understand the history of Québec nationalism.

First, cosmopolitans posit that nationalism and the partial preference towards the interests of conationals—individuals who mutually recognize one another as members of the same nation—over the interests of foreigners can be problematic due to its exclusive tendencies. They contend that nationalist attitudes tend to exclude some individuals who do not fit with the values of the majority culture of a particular nation (among the influential scholars, see especially Habermas 1998: 132; 2001, chapters 4 and 5; Caney 2005; Brock 2009; Tan 2004; Arneson 2016). Typically, cosmopolitan scholars argue that we do not have any special obligations towards conationals simply on the basis of sharing a common national identity. For them, we should first see ourselves as having equal obligations of justice towards all human beings, close or distant, as national ties are ultimately arbitrary and should not be used to exclude some individuals from considerations of justice.

What this position entails culturally for national belonging is most prominently illustrated by Habermas’s post-nationalist theory.Footnote 2 He argues that individuals living in a same state should not be encouraged to share particular values such as a shared history, a shared language, or be encouraged to see themselves as sharing a same descent, or a shared normative conception of the good life to allow for the proper functioning of a given society. Rather, they should only be encouraged to develop a shared identity organized around universalistic principles of human rights and the rule of law (Habermas 1998: 118; Leydet 2017). Hence, the baseline is that citizens should subscribe to a minimal shared identity that includes a shared commitment to democratic procedures, fundamental universal human rights and the rule of law. These commitments should be sufficient to ensure the social integration of all citizens of a democratic society (Habermas 2001). In other words, a certain cultural homogeneity, beyond this specific commitment to democratic values, is neither necessary nor desirable to ensure the integration of all citizens for Habermas. Following this approach, states might be instrumentally useful to protect the rights of individuals, but nationhood cannot explain why individuals should share some particular values or why citizens have particular obligations of justice towards one another.Footnote 3

Second, there are two nationalist approaches clearly opposed to the cosmopolitan one: the liberal nationalist and the communitarian position. However, the debate between liberals and communitarians, which received a great deal of attention around the turn of the millennium, is now less central in contemporary debates, which have shifted towards the more general question of what kinds of political and social relations justify the application of principles of justice between two individuals. However, they largely tend to ignore the question of what “special” features characterize nations.Footnote 4 As such, the main debates in contemporary political philosophy address the issue of what distinguishes relations between conationals from relations between conationals and foreigners rather than what distinguishes nations from other kinds of social groups. Yet, as we argue below, the concept of nation remains very important to make sense of individual identity—and then political decisions.

Despite their differences, liberal nationalists and communitarians both argue that a shared national identity is valuable and even necessary to allow for the integration of all citizens. The pre-existence of a common nationality, understood as either a shared culture, shared values, a shared conception of the good life or a commonly shared sense of identification to the same national identity, is a sine qua none to generate “desirable feelings of identity and belonging”, as put by Leydet (2017). The proper social and political inclusion of individuals can only be guaranteed through this shared national identity. However, liberal nationalists and communitarians differ on the extent to which the members of a nation should be inclusive and accommodating of cultural differences and minority cultures.

For instance, liberal nationalist authors such as Miller (1995, 2000, 2016) or Kymlicka (2003), Kymlicka and Norman (2000), claim that the concept of nationhood is neither outdated nor inherently problematic, although it has to be defined in an inclusive (and liberal) way. Miller argues for the importance of a shared public culture which can include not only political norms and principles (such as a basic commitment for democracy), but also social norms like “honesty in filling your tax returns” and cultural ideals such as “religious beliefs or a commitment to preserve the purity of the national language” (Miller 1995: 26).

Although a public culture is, for Miller, not necessarily all encompassing and should allow for significant individual leeway to develop particular “private cultures”, the common public culture necessary to ensure the effective social and political cooperation of all conationals cannot be reduced to a shared commitment to political and legal norms. Accordingly, for instance, Miller opens the door to the legitimacy of citizenship tests ensuring that immigrants have incentives to familiarize themselves with the host nation’s norms (both political and social) and culture. However, the author recognizes that a national identity is open to changes and may be flexible over time (Miller 2016).

Similarly, Kymlicka (2003), Kymlicka and Norman (2000) states that some things can be expected of individuals to ensure their social integration, although the nation state has to respect a multicultural approach which allows for individual and group variations within the nation. As such, even if nationalism remains valuable, we should adopt a “thin” understanding of nationhood. For instance, Kymlicka writes:

“citizens are still expected to speak a common national language, share a common national identity, feel loyalty to national institutions and share a commitment to maintaining the nation as a single, self-governing community into the indefinite future” (Kymlicka 2003: 273).

Accordingly, for him, some aspects of nationalism are necessary to allow for cooperation between individuals and it remains an important part of individual identity.

Other scholars associated with the nationalist school of thought have, however, developed a “thicker”, more substantial conception of nations. For these communitarians, it is not only seen as an important component of how individuals conceive of themselves and an important shared framework regulating their interactions, but nations are also understood as the legitimate expression of a group’s cultural identity. Following this communitarian conception, the very identity of individuals is grounded in the identity of the groupFootnote 5 (Seymour and Gosselin-Tapp, 2018: 34-40; Taylor, 1994). The nation is not only understood as a valuable part of an individual’s identity as a member of a collective enterprise, but the nation is a pre-existing condition to the very possibility of individual identity (MacIntyre 1994; Nielsen 1998; Lagerspetz 2000). This particular and fundamental role of the nation is clearly illustrated by Nielsen (1993: 32) in an article on Québec separatism:

“We are, to put it crudely, lost if we cannot identify ourselves with some part of an objective social reality: a nation, though not necessarily a state, with its distinctive tradition. What we find in people—and as deeply embedded as the need to develop their talents—is the need to be able not only to say what they can do, but to say who they are. This is found, not created, and is found in the identification with others in a shared culture based on nationality or race or religion or some slice or amalgam thereof. Given this aspect of our human nature, national consciousness and the forging and sustaining of a nation are extremely important to us whoever we are. Under modern conditions, this securing and nourishing of a national consciousness can only be achieved with a nation state that corresponds to that national consciousness”.

In other words, for communitarian authors such as Nielsen, belonging to a nation is a necessary component allowing for anything like an individual identity. Given this fundamental role of nationality, this thicker conception of nations can justify not only the protection of some minimal political values and cultural ideals (like the preservation of a common language or democratic values), but it can also be used to legitimize the protection of some particular ways of life even if they can sometimes infringe on universal human rights and liberal policies to some extent—such as freedom of religion or gender equality (Rawls 1999: 62–67).

Of course, as transpired in the above Nielsen’s quote, the case of Québec (and of Canada) is particularly interesting to study nationalism. Québec is both a national majority, which controls its own (provincial) state, and a French-speaking minority within the larger Canadian and North American context. Unsurprisingly, communitarian authors such as Nielsen highlight Québec’s fundamental right to secede from Canada if it wished to do so to affirm its distinctive national identity. However, he is prudent on the question of minority rights and maintains that a group right to self-determination is only legitimate to the extent that national minorities (e.g. anglophones in Québec) enjoy strong protections from majority decisions (Nielsen 1993, p. 30).

Communitarian approaches to nationhood, with their focus on the priority of the nation over individual identity, have important consequences on how individuals themselves do or rather should approach their own identity. As will be detailed below, communitarian approaches, with their insistence on how the national identity is found in the identity of the group (i.e. given as is) rather than constructed by the individuals themselves, thus defend ethnic definitions of nations. Though the meaning of “ethnic” is polysemous, it is here understood in terms of ascriptive criterion such as language, race and/or religion. Accordingly, communitarians in Québec typically focus on the fundamental importance of the French language, the shared French Ancestry of Quebeckers or the Roman Catholic roots of Québec’s culture (Stevenson (2006: 10)—see also Siegfried 1906). Given that the group identity precedes individual identity, it should be possible to circumscribe the nation by insisting on shared group characteristics.

In contrast, cosmopolitans tend to adopt a purely attainable conception of the state where nationalism should be inclusive of all those who live permanently on a given territory and are willing to cooperate with others. Following this approach, a certain individual commitment to the political institutions of a given country coupled with a sufficient devotion to universal human rights should be sufficient to be considered a part of a given state. Interestingly, Canada is sometimes presented as a cosmopolitan success story. Some authors have highlighted that though national identities remain important in the Canadian context and despite its participation in “imperialistic ventures” (Berger 1970, 259), it is particularly hospitable to cosmopolitanism because it is characterized by important ethnic and linguistic diversity (Kymlicka and Walter 2012: 12–13; Franck 2006: 37; Brysk 2009: 93; Welsh 2004: 189).

Similarly, Québec was also described by some as a civic nation which does not necessarily rely on a fundamentally ethnic conception of nationhood. Though Québec possesses a national language, French, some have underlined Québec’s “Americanity” (américanité)—see notably Thériault (2005, 2012) for a historical reading and critique of this term. It highlights the idea that Québec adheres to a broad North American culture defined by individualism, democracy, openness and civic nationhood with French being its only distinct ethnic characteristic (Bouchard 2000).

Finally, liberal nationalists distinguish themselves from both the communitarian and from the cosmopolitan reading of Québec’s culture by insisting on the importance of both attainable and certain ascriptive conditions. For instance, scholars such as Miller or Kymlicka both consider Québec nationalism to be a “nested” identity within Canada’s own national identity (Miller 2016: 143; Kymlicka 2003, 283–7). This entails that different national identities are not necessarily inconsistent with one another, but might diverge on the content of their public culture. Accordingly, they can recognize that Québec’s national culture can be defined by a commitment to preserve the French language and to some cultural or social norms or even some “tribal associations” which can remain important to distinguish the Québec nation from other national groups like Canada and the USA (Thériault 2012, 70). However, as mentioned above, they both insist that nations should also allow for significant individual leeway within the national culture.

From theory to practice

While highly (if not exclusively in some cases) abstract, these political theories are directly connected to fundamental empirical concerns. That is, not only are these questions normatively important in themselves to understand the value and importance of nationhood in the interactions between fellow citizens and between fellow citizens and foreigners, but they also affect issues of public policy. Further, beyond the theoretical and normative considerations supporting the three different schools of thought (cosmopolitanism, liberal nationalism and communitarianism) we argue that the citizens’ perspective on these questions is fundamental to fully grasp the practical importance of nations. Accordingly, our paper aims to make two empirical contributions: (1) we provide the first in-depth empirical examination of how Québec citizens typically conceive their own nation and (2) we examine how important this conception is in citizens’ vote choice calculus when it comes to casting a ballot such as in the Québec 2018 election.

In social sciences, the different definitions of nation belonging were historically defined following civic or ethnic criteria—see notably Kohn (1944) and Brubaker (1992)—which roughly mirror the difference between cosmopolitan and communitarian conceptions of the nation in philosophy as transpired at the end of the above section. However, conceptions of nationhood can highlight complex interrelations between these two poles. Notably, as the position of liberal nationalists shows, one can conceive of nationhood as a complex notion involving both civic and ethnic criteria. This more nuanced conception of nationhood which moved beyond the civic/ethnic dichotomy is also shared by some political scientists and sociologists—see notably Brubaker (2004), Goodman (2015), Wright (2011) and Bertosse and Duyvendak (2012). These studies highlighted to different degrees that national belonging and identification is subject to gradations (Goodman 2015; Wright 2011), to overlaps between nationhood and other group identifications (Brubaker 2004), and can be idiosyncratic depending on a nation’s particular history (Bertossi and Duyvendak 2012).Footnote 6 Accordingly, contemporary research in political science has moved more towards the question of how to measure each national public which can have their own image of what “traits and behaviours (sic.) are constitutive of national members”, as shown by Goodman and Alarian (2019: 6).Footnote 7 Some studies have also tried to develop post-nationalist models based on liberal democratic norms to provide alternatives to nationality-based explanations of citizenship (Soysal 1998; Goodman 2014).

Yet, the literature in both philosophy and political science tends to overlook how in-group definitions of the relevant national markers empirically inform individual behaviours and attitudes such as voting behaviours.Footnote 8 In this research, we provide such an empirical inquiry based on the fundamental theoretical debates outlined above and we also show that this new conception of nationalism allows us to make sense of electoral outcomes. To do so, we use the case of the Québec 2018 election, which we briefly describe below.

The Québec 2018 election—Beyond the ethnic/civic divide

As mentioned in the introduction, the Québec 2018 election is interesting for its historical value, notably because it ended almost half a century of bipartism between the PLQ and the PQ. It is also important to mention that the case of Québec nationalism in general was extensively studied in the 1990s and early 2000s to evaluate the legitimacy of secessionist claims (see notably Buchanan 1991, 1997; Miller 1997; Norman 2006; Nielsen 1993). Further, as mentioned, the case of the Québec nation is particularly interesting for studies of nationhood given its position as a majority, which controls its "national" (provincial) state within the Canadian federation, and as a French-speaking minority in the North American context.

This position as a French-speaking minority within the largely English-speaking North American context is observable in Québec by a continuous commitment to protect the French language and ensure its survival over time and by a certain opposition to Canadian multiculturalism. Québec’s history and how one should define its national culture are further a debated and contentious issue as transpired above when we considered the conflicting cosmopolitan, liberal and communitarian positions. Additionally, several studies have interestingly considered how the civic–ethnic dichotomy has historically played out in Québec in comparison to the Canadian context. Notably, Breton (1988) compared the evolution of nationalism in Québec and English Canada to show how both moved from ethnic to more civic conceptions of the nation, although in slightly different ways. In contrast, Riggs (2002: 40) has argued that ethnic nationalism should be understood as a more recent phenomenon on the part of some nations in reaction to increasing globalization.

However, as argued above, this dichotomization between civic and ethnic nationalism, where civic nationalism is often seen as the liberal and inclusive collective project, while ethnic nationalism is seen as exclusive or reactionary, misses the more complex connections between different civic and ethnic markers. Interestingly, some authors have argued that Québec nationalism is not necessarily exclusive, but can in fact be seen as a kind of nationalism which moves beyond the civic–ethnic dichotomy (Blad and Couton 2009).

As a matter of fact, different governments in Québec have asserted the desire to develop an “intercultural” model to approach the integration of newcomers within the Québec nation which is not necessarily exclusive or illiberal, though it remains ethnic to a certain degree (Gagnon and Iacovino 2016: 113; Rocher et al. 1995; Bouchard 2016; Anctil 2016; Seymour and Gosselin-Tapp 2018: 157–160). This intercultural model, compared to Canadian multiculturalism, is typically characterized by its greater insistence on the protection of the Québec majority culture (while Canadian multiculturalism is typically characterized by the rejection of the existence of one dominant Canadian culture) and the promotion and protection of a common public language (a concern largely absent from Canadian multiculturalism). It is also characterized by an emphasis on the integration of immigrants to the fundamental values entrenched in Québec culture (such as gender equality and democratic norms) (Bouchard 2012: 99–105).

Additionally, previous studies have highlighted the impact of Québec nationalism on the party system and the structuring strength of the issue of independence on political behaviour (for recent examples, see Bélanger and Nadeau 2009; Bélanger et al. 2018; Daoust and Dassonneville 2018; Nadeau and Bélanger 2013). All these characteristics of the Québec nation underline how this case study is at least presumptively interesting for considering how nationalism is conceived of by the citizens, how it can push our understanding of the possible relation between civic and ethnic aspects of nationhood, and how it might influence individual behaviour.

The 2018 election in Québec is also interestingly characterized by no concrete threat of a referendum on Québec independence, as the PQ leader pledged not to hold such as referendum in a first term if elected. Nationalism was, however, very present throughout the campaign. Among other things, important concerns were raised about the protection of Québec culture, notably prompted by issues surrounding the secularism of Québec’s public institutions and issues surrounding immigration. The winning party, the Coalition Avenir Québec (CAQ), promised to ban the wearing of all religious symbols for public employees in a position of authority (including teachers) and to reduce the number of new immigrants by 20% to better respect Québec’s capacity to integrate newcomers to Québec’s majority culture. Additionally, the CAQ promised to establish a mandatory test to evaluate the newcomers’ knowledge of the French language and of Québec’s fundamental values. Succeeding on this test is supposed to be a necessary condition for the issuance of a certificate of selection that is itself necessary for a newcomer to become a Canadian citizen.

So far, little research has been done to make sense of this election’s outcome. It is likely that publications will come out in the coming years (and the current special issue is part of this scientific effort). However, no research has provided a deep examination of Québec citizens’ conception of the nation and of its electoral consequences. We aim to do so, and we now turn to the empirical strategy.

Empirical strategy: data and indicators

In order to examine citizens’ view on what makes an individual part of the nation and its impact on the Québec 2018 election, we use a post-electoral survey conducted during the 3 weeks (October 10–30th) following the election by the polling firm Léger Marketing. The sample consists of approximately 3000 respondents who were randomly recruited from Léger’s web panel, and we weigh these data according to census data for age, gender, education, region and language. However, as we will look at the impact of nationalism on vote choice, we excluded citizens who were not eligible to vote. (There were about 250 sixteen- and seventeen-year-old respondents in the original sample.)

We measure ethnic nationalism using respondent’s opinion on the importance of certain criteria of what it means to be “Quebecker”, that is, to be part of the Québec nation. The exact question was as follows: “Some people say that the following things are important for being truly Québécois. Others say that they are not important. How important do you think each of the following is?” There were some criteria clearly related to civic nationalism—respect laws and institutions, feeling Québécois, to share Quebeckers’ values. However, these criteria, tapping a cosmopolitans’ view of nationalism, display very little variance and are not very insightful in that sense. For example, a vast majority agree that it is very important (62.8%) or somewhat important (28.2%) to respect Québec’s laws and institutions. As such, considering that the cosmopolitan focus on civic aspects of national identity does not appear to be particularly contentious, we focus on the more (ethnic) nationalist criteria which corresponds to the more liberal or communitarian view. These criteria are: to have been born in Québec, to have lived in Québec, to be able to speak French, to be a Catholic and to have French ancestry. Another important feature of ethnic nationalism is the in-group preferences or “tribal associations” to reuse Thériault’s (2012) vocabulary. Hence, we also add to the five indicators a measure capturing Quebeckers’ opinions on immigrants’ contribution to the Québec and more specifically on how they agree with the following statement: “Immigrants make an important contribution to Québec”. The options were strongly agree, somewhat agree, somewhat disagree and strongly agree. Attitudes about immigration’s contribution to Québec very likely measures deeply rooted values on nationalism and the conception of who is a full “contributor” (i.e. part of) to the nation.

To make sure that the more ethnic criteria, combined with the in-group preferences, tap into one single dimension (of ethnic nationalism), we ran a latent class analysis (LCA) on all the items and computing different number of clusters. We provide an overview of that method in the Supplementary Material just before Table SM.1. In this table, we display the indicators of model fit and Table SM2 shows the results for the two cluster (latent class) model. Overall, there is a probability of 0.57 for a respondent to be classified closer to the first factor (civic nationalism) and 0.43 for the second factor (ethnic nationalism). Clearly, those associated with ethnic nationalism were much more prompted to believe that all of the criteria were important.Footnote 9 However, the increase is much more important for the variables linked to a more liberal or cosmopolitan perspective compared to a cosmopolitan one. These are the first six variables shown in Table SM1, intuitively associated with a more ethnic view of the nation. Hence, we used the five criteria (born in Québec, to have lived in Québec, to be able to speak French, to be a Catholic and to have French ancestry) and the in-group preferences to create an index of ethnic nationalism. Empirically, the internal consistency of these criteria is very high as confirmed by the Cronbach’s alpha of 0.74, also strongly suggesting that they measure one single factor (Table 1).

Table 1 Ethnic nationalism indicators
Full size table

To examine who is more likely to have a lower/higher score value on this ethnic nationalism scale, we use different sociodemographic variables as well as underlying values and beliefs. The sociodemographics are age, gender, education (completed high school; post-secondary education but no university degree; university degree), language, place of birth and religious affiliation.Footnote 10 We argue that these are all causally prior to the formation of citizens’ nationalist attitudes. The underlying values and beliefs are: support for Québec independence and left–right ideology.

To analyse the impact of nationalism on vote choice, we measure the latter with a simple question asking whether the respondent voted and if so, for which party.Footnote 11 The distribution of this variable is quite close to the actual vote shares (in our survey, the CAQ receives 37.7%, the PLQ 24.5%, the PQ 20.6% and QS 17.1%).Footnote 12 To analyse vote choice, we follow Daoust and Jabbour (forthcoming) and use multinomial logistic regression as the variable is categorical. In order to control for factors potentially related to both ethnic nationalism and vote choice, we reduce the possibility of omitted variable bias by including important covariates such as age, gender, education, language, religious affiliation, place of birth, support for Québec independence and left–right ideology. The exact question wording for all the variables that we use is given in Supplementary Material.

Results

The fundamental importance of (some) ethnic factors

As mentioned above, many studies have considered how different conceptions of nationalism affect individual support towards multicultural and immigration policies (see note 7 above). However, to our knowledge, no other empirical study defines ethnic nationalism in the way proposed in this paper. Since we have little comparative evidence, we begin with a thorough description of citizens’ attitudes towards the nation. More specifically, we first show the mean of all five criteria and the in-group preference used to construct our ethnic nationalism index, where a greater value indicates a view that is closer to an ethnic conception of the nation.

There is quite a lot of variance in the means, but according to citizens, being Catholic and having French ancestry are much less important considerations to be part of the nation. That is, only 3.7% said that being Catholic was very important and 7.5% said the same for having French ancestry. To the contrary, “having lived in Québec” and “being able to speak French” are perceived to be the most important factors with means of 0.64 and 0.85, which is very high considering that our variables range from 0 to 1. Furthermore, the value regarding immigrants’ contributions indicates that a clear majority of respondents were positive (i.e. agreed that their contribution is important); that is, they displayed an in-group preference that is not very important.

Taken together, these indicators form our ethnic nationalism index and we show the distribution of this index in Fig. 1. The most striking feature is that it constitutes roughly a normal distribution (at least, much more than a polarized binominal one). Furthermore, the theoretical literature on cosmopolitan, liberal and communitarian views of the nation, and our empirical tests, also allow us to justify our choice of combining these indicators. This also supports the conclusion that a purely cosmopolitan approach to Québec’s national culture which would focus on its “americanity” is not sufficient to fully grasp the subtleties and different dimensions of this national identity. Thus, we need both civic and ethnic markers to approach and understand the Québec nation.

Fig. 1
figure 1

Distribution of ethnic nationalism. Notes N = 1878

Full size image

What predicts ethnic nationalism?

But who obtains greater scores on this nationalism index? Are there characteristics that systematically correlate with having a lower/greater score? Figure 2 shows the results of an OLS regression. Model 1 only includes the sociodemographics and Model 2 adds a set of underlying beliefs and values, which are causally posterior. As the displayed values are OLS regression coefficients and our dependent variable (ethnic nationalism) ranges from 0 to 1, Fig. 2 provides the direct effect of each variable.

Fig. 2
figure 2

Determinants of ethnic nationalism. Notes OLS regression coefficients with 95% confidence intervals. N = 1878

Full size image

The sociodemographic variables in Model 1 display interesting effects. First, it is surprising that variables such as age, gender and language, which usually predict views on nationalism, do not reach the conventional level of statistical significance in this case. Even when a variable does reach statistical significance, such as religious affiliation, the impact is very modest. An individual who is affiliated with a religion is predicted to have a score of ethnic nationalism that is 0.02 greater than a respondent who is not affiliated to a religion. In other words, it reaches statistical significance, but the impact is very small.

Second, two variables stand out as having a non-trivial impact. On the one hand, being born in Québec is associated with scoring higher values on our ethnic nationalism scale. This effect is of approximately + 0.05. While the positive effect is not very surprising, the magnitude should not be overstated and is still quite modest. On the other hand, education displays a negative impact of − 0.14, showing that the most educated have lower scores on our ethnic nationalism index. This maximum effect reflects someone who, ceteris paribus, moves from having “completed high school” to having “completed university”. But even an increase of one category (out of three in total), such as moving from “completed high school” to “post-secondary education but no university”, results in a − 0.07 effect (i.e. half of the maximum impact). This is not very surprising as the more educated are known to display more liberal attitudes in general, but the magnitude of the effect of ethnic nationalism is very interesting and stands as an outlier among the sociodemographics.

Model 2, shown in Fig. 2, includes respondents’ view on Québec independence and left–right ideology. Although being in favour of Québec independence is associated with scoring greater values on our ethnic nationalism scale, the effect is very modest. That is, respondents who would vote yes to a referendum on Québec independence display on average a score 0.03 greater than those who would not vote yes. However, ideology stands out with a larger coefficient, but that variable’s scale makes the comparison difficult. For a clearer assessment, we must take into account that the range of this variable was measured on a 0–10 (rescaled 0–1) scale. Hence, the coefficient of + 0.12 represents an important maximum effect, but a more conservative evaluation could be to compare someone located at 0.3 (one standard deviation below the mean) on the ideology scale to someone located at 0.7 (one standard deviation above the mean), for an impact of 0.04. All in all, underlying values and beliefs are related to ethnic nationalism, but this effect is not very strong.

Ethnic nationalism and vote choice calculus

We now turn to the impact of ethnic nationalism on citizens’ vote choice calculus. More specifically, does ethnic nationalism correlate with support for some specific parties and desertion of others? Full results of the multinomial regression can be found in Table SM2, and Fig. 3 shows the predicted probabilities to vote for each party as a function of the score on the ethnic nationalism scale.

Fig. 3
figure 3

Impact of ethnic nationalism on vote choice. Notes Predicted probabilities with 95% confidence intervals are shown. Based on regression model in Table SM2

Full size image

Whether an effect is considered statistically significant or not depends on the reference category, so we focus on the substantial effects. Three of the four parties’ support seems to be affected by voters’ degree of ethnic nationalism. The CAQ is the party for which the effect is most interesting. Someone at 0 has, ceteris paribus, a predicted probability of 0.2 to support the party, while it increases to almost .6 at the maximum score of ethnic nationalism. A more realistic comparison could be the total effect from one standard deviation around the mean, which consists of comparing someone at 0.25 to 0.65 since the mean is 0.45 and the standard deviation is of 0.2. The effect is, even in that context, of 15 percentage points, which is quite important. This seems to be in line with the public discourse around the Coalition Avenir Québec and their more restrictive pledges on immigration, which were coherent with the nationalist and communitarian view of a commitment from immigrants that goes beyond the minimal approach of cosmopolitans. However, this is somewhat surprising giving that the party is supposed to be a "coalition" focused on the economy, which was fundamental when François Legault created the Coalition Avenir Québec with federalists, independentists, progressives and conservatives.

The effect of ethnic nationalism is less important, but still noteworthy for the Libéral Party of Québec and Québec Solidaire. The size of the negative effect is very similar in both cases, with a maximum impact of about 15 percentage points. This is not surprising as the two parties, even if they differ on their official view on Québec independence, are strongly committed to multiculturalism and are much closer to a cosmopolitan view of the nation compared to the Parti Québécois or the Coalition Avenir Québec.

The PQ did not benefit or lose from more ethnic nationalism. Surprisingly, the effect is remarkably flat. As mentioned, the party rejects multiculturalism and cosmopolitanism views of nationhood, and, most importantly, this was historically demonstrated in its public policies. Among other things, the PQ had an important historical role in the protection of the French language in Québec through the adoption of the Charter of the French Language (also known as “Bill 101”) which defines French as the official language of the province. It is also important to mention that the PQ also proposed a secularization bill in 2013, favouring the religious neutrality of all employees of the state by forcing anyone giving or receiving a service from a provincial governmental agency to uncover their face. Further, in 2018, the PQ also presented a number of nationalistic pledges in its platform including an emphasis on the francization of immigrants and refugees and the application of Bill 101 to small enterprises and federal organizations. All in all, these null results for the PQ although less impressive at first glance are proved to be quite interesting.

Discussion

How do Quebeckers conceive their own national identity and does this influence their vote choice calculus? The particularities of Québec nationalism, with its importance of a shared public language and the integration of newcomers to the fundamental values shared by the majority, provide interesting inputs on how we should understand the practical and normative importance of nationalism. While the theoretical debates reviewed are very insightful, we aimed to move beyond theory in order to tackle the empirical reality of Québec society. As discussed above, it does appear that some aspects of ethnic nationalism are important to explain how Quebeckers understand their own identity, which raises important challenges for cosmopolitans and communitarians alike or for authors who distinguish strongly between ascriptive and achievable traits to define nations.

First, cosmopolitans who try to develop a post-nationalist conception of citizenship and social and political integration could still try and deny that ethnicity (as understood in this paper) is a necessary component to ensure the proper integration of immigrants within a particular nation or that it is non-instrumentally valuable. However, it would then be necessary to provide an explanation of how there can be a disconnection between what individuals value as important aspects of their own individual identity with respect to nationhood and the proper social and political integration of newcomers.Footnote 13 This appears to be a rather challenging task.

Second, our empirical results also address a strong challenge to communitarians who consider that nations are valuable in and of themselves and are delimited by ascriptive ethnic factors that lie outside of individual control (such as one’s ancestry, one’s place of birth and one’s religious affiliationFootnote 14). Rather, it appears that some more or less achievable traits are predominant to explain nationhood in Québec: having lived in Québec and, most importantly, the ability to speak French. In a manner akin to the challenge addressed to cosmopolitans, authors who want to argue that belonging to one nation is prior to one’s individual identity and precedes individual autonomy should thus provide an explanation for why, at least in the particular case of Québec, national belonging seems achievable even for most respondents who adopt a more ethnic view of the nation.

Overall, we believe that our research offers three main contributions. First, it builds bridges between political theory and electoral studies, which too often ignore one another. Second, with our focus on the Québec nation, our study provides novel information to understand the impact of nationalism on individual behaviour. Two aspects of our results are worth highlighting in closing. Firstly, it is interesting to note that the distribution of ethnic nationalism is more "normal" than polarized in an era where the conventional wisdom seems to be that polarization is on the rise. Secondly, our results further illuminate the relevant predictors to assess whether or not someone is more or less likely to have an ethnic conception of the nation. That is, education and ideology are the two somewhat influential predictors, while the others (age, religion, gender, language, etc.) only have, at best, a significant (but substantially modest) impact on one’s position in our ethnic nationalism continuum.

Our third and final main contribution is to improve our understanding of the 2018 election in Québec by highlighting the impact of nationalism on voter’s choices. Though it might prima facie seem that nationalism was of lesser importance considering that political independence was not a predominant issue (the PQ having pledged not to hold a referendum on the issue), we have shown that nationalism is still important and its impact on voters’ behaviour goes beyond the question of national independence. In fact, the impact was very important for the CAQ, which benefited from voters holding more ethnic nationalist attitudes, while there was surprisingly no effect for the PQ. Moreover, there was a less important but still noteworthy negative effect of ethnic nationalism and voters’ proclivity to support the Parti liberal du Québec and Québec Solidaire.