One Cultural Cleavage and Two Cultural Value Divides?

Cleavages represent structurally embedded social groups with contrasting interests and values that inform distinctive voting preferences (Lipset and Rokkan Reference Lipset, Rokkan, Martin Lipset and Rokkan1967; Bartolini and Mair Reference Bartolini and Mair1990). Among the cleavages outlined by Lipset and Rokkan in their book Party Systems and Voter Alignments: Cross-national Perspectives (1967), the class cleavage captured most attention of researchers in the subsequent decades. It refers to the class-based opposition between Left and Right parties, representing opposite stances on the ideal relationships between state and market (Lijphart Reference Lijphart1982; Dalton Reference Dalton and LeDuc1996).

However, class interests were not the only factor that influenced voting behavior and party systems, as Lipset and Rokkan also acknowledged various other cleavages, like the one based on religion (Lipset and Rokkan Reference Lipset, Rokkan, Martin Lipset and Rokkan1967). Across Western Europe, this religious cleavage either overlapped with other societal divides like in Norway (Ertman Reference Ertman, Van Kersbergen and Manow2009) or constituted an independent dimension of political competition, e.g., via parties of religious defense in countries like Belgium and the Netherlands (Kalyvas Reference Kalyvas1996). It reflected the conflict between church and state, or, at the individual level, between religious and secular individuals, with the religious more likely to vote for Christian-Democratic (or other distinctively culturally conservative) parties (Knutsen Reference Knutsen2010). At the heart of this cleavage lies a value divide about the societal role of religion, i.e., the degree to which religious norms should be reflected in state laws and followed by citizens.

Although Western party systems at the beginning of the 1960s still reflected the “frozen” social cleavages of the 1920s (Lipset and Rokkan Reference Lipset, Rokkan, Martin Lipset and Rokkan1967), the next decades were marked by far-reaching changes. In particular, the so-called “counter culture” of the 1960s and 1970s brought cultural issues pertaining to individual liberty to the political agendas of Western European countries (Marwick Reference Marwick1998; Campbell Reference Campbell2007; Houtman, Aupers, and De Koster Reference Houtman, Aupers and De Koster2011). These changes sparked academic debates about the evolution of cleavage-based politics, with special attention to the emergence of a so-called “new cultural cleavage” that has allegedly only gained in social and political salience since (Inglehart Reference Inglehart1977; Kriesi Reference Kriesi1998; Reference Kriesi2010).

These debates have produced various characterizations of the cultural value divide held to be central to the “new” cultural cleavage, e.g., “libertarian/authoritarian” (Kitschelt and McGann Reference Kitschelt and McGann1997; Flanagan and Lee Reference Flanagan and Lee2003; Stubager Reference Stubager2008), “post-materialist/materialist” (Inglehart Reference Inglehart1977), “self-expression/survival” (Inglehart and Baker Reference Inglehart and Baker2000; Inglehart and Welzel Reference Inglehart and Welzel2005), “green-alternative-libertarian/traditional-authoritarian-nationalist” (Hooghe, Marks, and Wilson Reference Hooghe, Marks and Wilson2002), and “libertarian-universalistic/traditionalist-communitarian” (Bornschier Reference Bornschier2010). Central to these various definitions of the cultural divide is the idea that cultural issues pertain to problems of social order and represent models for dealing with those problems. At least two of such models, one religious and one secular, are seen as underlying the value conflicts at the heart of the cultural cleavage (see for instance Norris and Inglehart Reference Norris and Inglehart2019). The religious model gives rise to conflicts between moral traditionalism and moral progressiveness, while the secular model gives rise to conflicts between authoritarianism and libertarianism.

The secular model of social order pits the authoritarian low educated and the libertarian highly educated against each other in their attitudes towards immigration and cultural diversity, law and order (Houtman Reference Houtman2003; De Koster and Van der Waal Reference De Koster and Van der Waal2007; Houtman, Achterberg, and Van der Waal Reference Houtman, Achterberg and Van der Waal2011). These issues pertain to cultural hierarchies and to (un)willingness to accept cultural diversity, not least the cultural diversity brought by immigrants in recent decades (Stenner Reference Stenner2005; De Koster and Van der Waal Reference De Koster and Van der Waal2007). The less educated are more likely to value sameness and conformity over diversity, and thus embrace authoritarian stances towards immigration, ethnic diversity, and law and order. The more educated, on the other hand, are more inclined to be libertarian and to embrace diversity and individualism (Houtman Reference Houtman2003; Van de Werfhorst and De Graaf Reference Van de Werfhorst and De Graaf2004; Stubager Reference Stubager2008; Reference Stubager2010).

The religious model of social order sparks conflicts between the religious and the secular about the legitimacy of traditional Christian doctrines (for the case of Western Europe) and behavioral norms (Finke and Adamczyk Reference Finke and Adamczyk2008; Storm Reference Storm2016). The latter pertains especially to matters of sexuality, life and procreation, family issues, and gender roles. Religious individuals understand Christian norms and guidelines on how to live properly as pre-given by a higher divine authority, as having proven their efficacy over centuries, and as distinct from and superior to conventional and man-induced secular laws (McCullough and Willoughby Reference McCullough and Willoughby2009). Non-religious moral progressivists, on the other hand, do not ground their moral principles in religion and reject the latter's claims of unquestionable authority and predefined social roles to be followed literally (Brown Reference Brown2009; Houtman, Aupers, and De Koster Reference Houtman, Aupers and De Koster2011).

These two cultural value divides, i.e., moral traditionalism-progressiveness and authoritarianism-libertarianism, have often been understood as together forming the heart of the “new” cultural cleavage. All things considered though, this is somewhat odd. In the first place, the moral traditionalism-progressiveness value divide is of course not “new” at all. It has in fact always been central to the religious cleavage as already discussed by Lipset and Rokkan (Reference Lipset, Rokkan, Martin Lipset and Rokkan1967) in the 1960s, well before the emergence of the “new” cultural cleavage. Moreover, with the rise of the “new” cultural cleavage secular issues pertaining to immigration and cultural diversity, law and order have become increasingly important (Kriesi Reference Kriesi2010), while religion and traditional morality simultaneously have lost much of their former significance in Western Europe. Yet, the two cultural value divides have often been found to be empirically interconnected (e.g., Evans, Heath, and Lalljee Reference Evans, Heath and Lalljee1996; Flanagan and Lee Reference Flanagan and Lee2003) and have therefore been widely combined into one single scale to represent the “new” cultural dimension in politics (Achterberg and Houtman Reference Achterberg and Houtman2006; Munzert and Bauer Reference Munzert and Bauer2013; Norris and Inglehart Reference Norris and Inglehart2019).

Various studies have meanwhile nonetheless pointed out that moral traditionalists and authoritarians are by and large different, non-overlapping groups, suggesting the need to disentangle rather than combine the two value divides (Stenner Reference Stenner2005; Reference Stenner2009; Lakatos Reference Lakatos2015). Indeed, research by De Koster and Van der Waal (Reference De Koster and Van der Waal2007) suggests that the often-found positive relationship between the two scales is due to moral progressivists rejecting both moral traditionalism and authoritarianism in the name of personal liberty. They conclude from this that moral traditionalism and secular authoritarianism are different value divides that need to be kept apart and treated separately rather than being combined into one single “new” cultural divide, as so many have done. This conclusion assumes that the relationship between the two value divides is identical across contexts, but the fact that their data come from the Netherlands, one of the most secularized and morally progressive countries in the world (see Norris and Inglehart Reference Norris and Inglehart2004), may suggest otherwise. We, therefore, theorize in what follows that it is precisely secularization that has sparked a rejection of both moral traditionalism and authoritarianism in the name of personal liberty. The outcome of it is strong correlations between the two in largely secular and morally progressive contexts and weak or even absent ones in massively religious and morally traditional ones.

Secularization and Value Divides: Hypotheses

The 1960s witnessed massive and rapid processes of secularization, driven by critiques of religion and the churches as standing in the way of personal liberty, not least in matters of sexuality, life and procreation, family and gender roles (McLeod Reference McLeod2007; Brown Reference Brown2009). The result was a decline of religion's social significance and a shift from the dominance of religiously informed moral traditionalism to a growing importance of individual liberty and personal authenticity. The so-called “counter culture” back then did however not merely critique religion and moral traditionalism, but was critical of authority generally, irrespective of whether the latter was religiously or secularly based (Zijderveld Reference Zijderveld1970; Roszak Reference Roszak1972). The counter culture as such exemplifies how secularization stimulates a rejection in the name of personal liberty of moral traditionalism and secular authoritarianism alike. Secularization, and more specifically the turn to moral progressiveness that comes with it, thus appears to have strengthened the relationship between the value divides of religiously informed moral traditionalism-progressiveness and secular authoritarianism-libertarianism.

The turmoil of the 1960s and 1970s has meanwhile left its lasting imprint on Western European societies (Marwick Reference Marwick1998; Houtman, Aupers, and De Koster Reference Houtman, Aupers and De Koster2011), not least in the realm of politics (Inglehart Reference Inglehart1977; Dalton Reference Dalton and LeDuc1996; Elff Reference Elff2007). Still today, the heirs of the counter culture, i.e., the New-Leftist parties and their secular voters and sympathizers, are critical about traditional religious as well as secular rightist-authoritarian tendencies (e.g., Dolezal Reference Dolezal2010; Kriesi Reference Kriesi2010; Stubager Reference Stubager2010). The latter tendencies have become much stronger due to an increased salience of issues of immigration and cultural diversity, law and order from the 1980s onwards (Bornschier Reference Bornschier2010). Especially immigration issues have since then conquered Western-European political agendas, with the most economically developed and culturally progressive Northern countries having become most visibly divided over these issues (Van Oorschot Reference Van Oorschot2006; Stubager Reference Stubager2010; Silva Reference Silva2018; Dennison and Geddes Reference Dennison and Geddes2019). This period since the 1980s has witnessed the emergence and electoral success of New-Rightist political parties that politicized these issues, not least by framing immigration as posing a threat to national identity and cultural sameness (Ignazi Reference Ignazi2003; Betz and Johnson Reference Betz and Johnson2004; Kriesi Reference Kriesi2008; Reference Kriesi2010; Dolezal Reference Dolezal2010).

This rightist-authoritarian backlash since the 1980s has in the process underscored how the 1960s have dealt more of a blow to religious moral traditionalism than to secular authoritarianism. While in the 1970s, advocates of personal liberty still primarily faced religious and morally traditionalist adversaries, from the 1980s onwards their enemies increasingly tended to have exclusively secular authoritarian profiles. Since the 1960s, then, religion and moral traditionalism have not only massively declined but have in the process also lost much of their former political salience, giving way to a redefinition of problems of social order in a secular direction (Akkerman Reference Akkerman2005; Achterberg Reference Achterberg2009). Yet, it is clear that religion and moral traditionalism have not disappeared in even the most secularized Western-European countries (Halman and Draulans Reference Halman and Draulans2006) and it would be similarly wrong to assert that the authoritarianism-libertarianism divide did not exist before the 1980s (see for instance Lipset (Reference Lipset1959) on working-class authoritarianism).

While the two cultural value divides have hence always co-existed, processes of secularization appear to have sparked a quest for personal liberty that has led rejections of moral traditionalism and of authoritarianism to increasingly coincide, producing stronger correlations between scales measuring the two. While this makes it understandable why so many scholars of contemporary politics have treated the two value divides as together constituting one new value cleavage (Evans, Heath, and Lalljee Reference Evans, Heath and Lalljee1996; De Witte and Billiet Reference De Witte and Billiet1999; Houtman Reference Houtman2001; Flanagan and Lee Reference Flanagan and Lee2003; Achterberg Reference Achterberg2006), it is then in fact not so much moral traditionalism and authoritarianism that have come to coincide, but rather their rejections by those foregrounding values of personal liberty. If this is indeed what has happened, the strength of the relationship between the two cultural value divides should be stronger in contexts where religion is less prevalent, because processes of secularization there resulted in a sizeable group of population who reject both traditionalism and authoritarianism.

If this theory about value change holds any water, it should be possible to confirm four hypotheses that can be derived from it. The first one predicts that the relationship between moral traditionalism-progressiveness and authoritarianism-libertarianism has grown stronger in time (H1). The three remaining ones address the explanation of this process. The second and third hypotheses predict that the link between the two value divides is stronger in more secularized contexts than in more religious ones (H2), because the latter are more morally progressive and less morally traditionalist (H3). This means that we expect lower levels of contextual religiosity to strengthen the relationship between the two value divides due to lower levels of contextual moral traditionalism. Finally, our fourth and final hypothesis predicts that the strengthening of the relationship between the two value divides across time (see H1) is caused by processes associated with secularization (see H2 and H3) and the effect of time is explained away by it (H4).

Data

We aim to study whether the link between authoritarianism and traditionalism has become stronger in time, due to a process of secularization that has increased moral progressiveness. To do so, we analyze the data of the European Values Study for 17 Western-European countries that all have a Christian heritage and not lived under Communist rule (EVS 2011). We use all four available waves of the study (1981, 1990, 1999, and 2008). The resulting dataset includes more than 69,000 individuals nested within 64 contexts, four waves, and 17 countries. Those countries include Austria (only three waves), Belgium, Denmark, Finland (only three waves), France, Germany, Iceland, Ireland, Italy, Malta, Netherlands, Norway (only three waves), Portugal (only three waves), Spain, Sweden, UK, and Northern Ireland.Footnote ¹ All descriptive statistics are available in the Appendix.

Measurement

In order to compute our dependent variable—a measure of connectedness between the moral traditionalism-progressiveness and secular authoritarianism-libertarianism value divides within each context—we first require individual scores for both divides and justify their use as separate value scales. The moral traditionalism-progressiveness scale measures respondent's moral stances on matters of life and death, procreation and family life, and is based on five questions indicating whether a respondent finds homosexuality, abortion, euthanasia, divorce, and suicide justifiable.Footnote ² For those who responded validly to at least four of these five questions, scale scores were assigned as mean standardized scores and then transformed to range from 0 to 10, with 10 indicating strongest moral traditionalism. The scale is highly reliable with an overall Cronbach's α of 0.82.Footnote ³

The measurement of authoritarianism-libertarianism posed more difficulties due to a scarcity of suitable items, but eventually, we chose for the best available option by selecting five questions; four about opposition to having (1) immigrants, (2) people of different race, (3) Muslims, and (4) ex-criminals as neighbors, and (5) one about whether they think the native-born should have priority in getting a job.Footnote ⁴ Because the questions about opposition to Muslims as neighbors and about privileging the native-born on the labor market were not asked in the first wave, scale scores were assigned as mean standardized scores to all those who responded validly to at least three of these five questions. The scores were then transformed to range from 0 to 10, with 10 indicating strongest authoritarianism. The resulting scale is fairly reliable with an overall Cronbach's α of 0.67.Footnote ⁵

Factor analysis with Oblimin rotation demonstrates that all authoritarianism items load on the first factor and all traditionalism items on the second. The two factors have Eigenvalues of respectively, 3.53 and 1.96 (before rotation), explain 90% of the variance together, and have a correlation of 0.24 (see Table 1).

Table 1. Moral traditionalism and secular authoritarianism items: factor analysis results

Blanks represent loadings <0.3; factor analysis (IPF) based on polychoric correlation matrix with Oblimin rotation; all items are standardized.

As indicated above, the correlation between moral traditionalism and authoritarianism is then computed for each of the country-year combinations separately to obtain the dependent variable to be used in this study. For all contexts combined the correlation is quite low reaching 0.22, the highest one being 0.29 for the Netherlands in 1990 and the lowest ones being statistically indistinguishable from zero (Northern Ireland in 1981 and Portugal in 1999). The intraclass correlation score of 0.34 suggests that multilevel modeling is required to account for country-level variation in the dependent variable.

The explanatory variables are (1) time, (2) contextual religiosity, and (3) contextual moral traditionalism. Time is measured as the wave of the EVS (1981–1990–1999–2008) and is introduced in the models as a continuous predictor. Contextual religiosity is computed for each of the country-year combinations as the mean score of an individual-level scale that consists of attending religious services at least once a month and believing in God, heaven, hell, sin, and life after death.Footnote ⁶ Although all these variables are binary, they are highly correlated and load heavily on one factor with an Eigenvalue of 3.21 that explains 89% of the variance. The resulting scale (mean standardized scores) is highly reliable with Cronbach's α of 0.86 and is recoded to range from 0 to 10 (most religious). The least religious context is Sweden in 2008 (2.33) and the most religious one is Malta in 1981 (9.28). Contextual moral traditionalism is computed for each of the country–year combinations as the mean score of the individual-level scale for moral traditionalism. The least and most morally traditional contexts are respectively, Sweden in 2008 (3.27) and Malta in 1981 (9.61). Contextual authoritarianism is computed in a similar fashion and used as a control variable in the statistical analyses. It ranges from 1.32 in Sweden in 1999 to 4.01 in Malta in 2008.