“Music for the Many”: A Quantitative Inquiry into Contextual Questions Surrounding National Anthem Adoption

This piece was published in the Winter Issue Print Edition (Volume 11)


National Anthems: Legislation Worth Exploring 

In his 1987 essay “On Being Alive to the Arts and Religion,” Trotter wrote that music “springs out of the very speech and soul of a person or a community.”[1] Music as the second art form is excellent at evoking feelings of ritualistic connection and lived communal identification, all the while serving as an accessible form of art to most audiences. National anthems have established a contemporary group ethos of music that various recognized international institutions have continued to standardize and reproduce since Argentina adopted the first continuous national anthem in 1816. Hummel finds that national anthems are among the most psychologically powerful state symbols insofar as they provide a means for countries to encourage their policy preferences and socialize their citizens.[2] Gilboa and Bodner corroborate these results in their study of the Israeli and Polish national anthems, reporting similar findings in their cases.[3]

National anthems serve as both tools for and intertextual representations of the state, demanding solvent opposition to qualify their existence. Yet this endeavor is not always successful, notably observed by the Japanese national anthem’s affixation to the state’s militarism rather than the Japanese nation itself.[4] Therefore, it is important to consider how anthems exist in a liminal state and are constantly being reconstructed and embodied by subnational communities. These symbols represent the tense and quasi-dialectic relationship constituting the nation and the state. Domestically, governments declare their anthems a cohesive representation of the nation; internationally, they must build the distinctiveness of their nation through a unique anthem. Anthems represent a complex synthesis of uniformity and individuality: countries must resist external influence on their anthem’s national character while also avoiding alienation of subnational groups. Yet anthems are not genealogies; relevant studies have not found former colonies to overwhelmingly adopt even similar anthems to those of their former colonizers.[5]

Recent research has found national anthems to be effective tools of national integration compared to other symbols, yet political scientists have registered scant interest in the research of national anthems comparatively.[6] For instance, a 2019 query of “national flags” in the JSTOR database results in 25,695 more journal articles than a similar search for “national anthems.” Such a research gap motivates individual countries to avoid modifications to national anthems as well; forty-five countries have changed their flag since 2000, only eighteen have since modified their anthems.[7]

Contemporaneously, the constructed utility of anthems as tools of the state have themselves become victim to circular logic. There is often a pervasive, non-partisan feeling within the internationalized paradigm that an anthem is necessary for a country to harmonize the balance of nation and state. In the case of contemporary England, both Labour and Conservative Members of Parliament have called for an official national anthem in the last 20 years—evidencing the perceived non-partisan nature of anthems in British discourse.[8] In addition, throughout the twentieth century, both revolutionary and reactionary popular movements adopted their own anthems through differing forms of top-bottom social relations as a means of conveying their valued “symbols, rituals, and traditions.”[9]

Yet, national anthems do not exist independent of the international community. Instead, they represent texts actively embodied and changed by the state and constituency. As nations discursively recast their anthems domestically, Adal articulates the simultaneous international reification of anthems, writing that “like national cuisines, national anthems cannot exist outside of a world of other national anthems. It is the international that opens up the space in which the national can exist. In other words, the national is necessarily mimetic.”[10]

This research interrogates the manufacturing of anthems as discourses, examining the relationship between an extant political history of fascism or authoritarianism, the level of cultural context, and the de facto or de jure constitutional secularity in relation to the time taken for a country to adopt a national anthem. To do so, I construct and examine anthologies of literature for each variable, operationalize and formulate hypotheses, and conduct a statistical analysis on the created data set. 

Theoretical Framework

Although the discipline of my research is established in discourse analyses and broader concepts of constructivism, neo-positivism lends a unique methodological approach to analyzing the incremental implementation of anthem adoption. Neo-positivist epistemology also acknowledges the importance of semiotics. Cunningham’s synthesis of semiotics and educational psychology indicates a shift in the semiotics community from an entirely interpretivist epistemology towards the inclusion of quasi-positivist “empirical constructivists.”[11] Representing the hardline interpretivists, Pressley et al. accuse Cunningham of being too forthcoming in his predictions of a mutual understanding connecting the psychological and semiotic communities. Yet, even they acknowledge that his methods on the inferential policymaking semiotics “[had] the potential for permitting rapid progress in understanding and testing the instructional beliefs and practices consistent with the semiotic process.”[12]

My research also utilizes the Soviet Tartu-Moscow School of neo-positivist semiotics by operating under the assumption that sovereign states collectively (either through mandate or suffrage) adopt communally understood symbols.[13] Although there are widespread differences in the mechanisms of adoption, the process of anthem adoption does not change the functional technocracy of national anthems. As argued by Mayo-Harp, “while national anthems are created by elites seeking to reach some specific goals, they are, at the same time, authentic expressions of popular identity.”[14] Therefore, with the exception of stateless nations (which lack territorial sovereignty, itself a key facet of Mayo-Harp’s argument), this study does not exclude cases on the basis of governance systems. 

Fascist Functionalism

Recent studies in semiotics have blurred the lines between national symbols serving as capable tools for multiculturalism or for fascist and authoritarian communities.[15] In particular, ethnomusicologists and political psychologists have studied the permeative effects of anthem components—such as tempo, lyrics, and tone—on native and non-native constituents. De Caprariis argues that Mussolini’s Partito Nazionale Fascista found the most success in group building through their anthem Giovenezza, composed with a similar cadence and tone to that of the socialist standard-bearer The Internationale.[16] Doğan critiques this liberal and fascist dichotomy of anthem forms, arguing instead that fascist parties necessarily adopt anthems independent of the state, thus eliminating the “fascist” form of anthem.[17] Using Giovinezza as a case, Doğan claims that fascist parties avoid affixing themselves to the state apparatus, preferring instead to establish their own private bureaucracy as a means of crowding out previous state organizations. Similarly, by 1934, Hitler had established separate Nazi institutions parallel to the existing German state bureaucracy— a labor union, athletic organizations, and think tanks (Ahnenerbe)—all under the umbrella of The National Socialist People’s Welfare (Nationalsozialistische Volkswohlfahrt). Doğan asserts that fascist parties rely on anthems in the same way, discouraging legislative bodies from adopting a state anthem in favor of instead popularizing their party’s anthem.[18] While Doğan intentionally limited her research to a case study of the Italian Giovenezza anthem instead of studying a broader range of fascist anthems, other semioticians have cited the Romanian Iron Guard party and their “Sfântă Tinerețe Legionară” anthem as an example of a fascist “unadopted anthem,” further triangulating Doğan’s theory.[19]

Opposite Doğan’s functionalist school of thought sits the post-structuralists. Erol, himself of the school, argues that scholars can only dissect fascist political parties after examining their socio-historical contexts.[20] Consequently, these scholars argue that national anthems are representative of distinct paradigms that match their relevant histories. Post-structuralism contradicts the functionalist claim that distaste towards national anthems is inherent to fascism; furthermore, several examples of fascism in Africa support this conviction.[21] I incorporate this conceptual disagreement into the research design by ensuring a rigorous coding process and maintaining elements of control between cases, here operationalized as the level of secularity. 

Authoritarianism as a Robustness Check

Epstein’s 2011 comparison of the biopic film genre to national anthems and critique of both as “superficial iconographic symbols” reveals the commonplace discourse surrounding national anthems in liberal democracies.[22]This uncertainty of anthem adoption, though operationally different from Doğan’s theory, produces the same outcome of fascist parties’ hesitancy to adopt anthems. Furthering this observation, Daughtry argues that the mechanism that leads to authoritarian regimes embracing national anthems is opposite to the hesitation held by fascist and liberal states.[23] Whereas fascist parties necessarily operate in exclusive and collective social organizations independent of the state apparatus, authoritarian regimes require state functions to preserve power, and thus are expected to swiftly adopt national symbols.

In their aforementioned case study, Gilboa and Bodner find that 84% of respondents recollect historic national events and 74% report experiencing an emotional feeling after listening to their national anthem—compared to 46% and 68%, reporting respectively, for the national emblem.[24] Bradley used similar results in her study of national anthems serving as hegemonic discourse, ultimately finding that the discourse over anthems within fascist parties and liberal states are similar. Methodologically, I use the mechanism for Bradley’s claim of anthem adoption in my operationalization of fascism: “The line separating a critical pedagogy [in deciding to adopt an anthem] based upon anti-racism principles and practices from the fascistic imposition is one that can easily be crossed…”[25] Bradley critiques Doğan’s theoretical mechanism preventing fascist parties from adopting an anthem, arguing instead that she could apply the same argument to liberal states.

Lughofer finds similar results in his biography of Paula von Preradović, the lyricist of the Austrian national anthem. Although Von Preradović demonstrated sympathies for the Austrofascist movement—and even composed the anthem with fascism in mind the postwar public never came to understand the anthem as inherently fascist.[26]Therefore, the anthem is still in place with minimal pushback, thereby displaying the similar perceptions of anthems compartmentalized as either liberal or fascist. Both Payne’s comparison between Iberian fascism and authoritarianism and Daughtry’s discourse analysis of the Russian national anthem find that the established similarities between fascist parties and authoritarian regimes are outside the breadth of Doğan’s theoretical reasoning.[27] Therefore, Bradley’s argument that anthems serve as an effective tool for isolated fascist parties is actually more applicable to authoritarian regimes in this research.

Constitutional Secularism

There exists a long-studied relationship linking the state’s de jure secularity and adopted forms of national symbols.[28] Meizel finds that religious symbols are effective tools for anthem framers, though there is no present consensus on whether religious regimes are more or less likely to adopt anthems than secular nations.[29] Following a reading of each available country’s constitution on the Constitute Project, I uncover two distinct binaural secular and non-secular categorizations.[30] In the first group, de jure non-secular states, such as Afghanistan and Ecuador, define their founding as one established in either the principles of a certain religious tradition or with the guidance of a specific religious organization. For an additional categorization, I expand the criterion to exclude any mention of a theist belief system that is coded as de facto secular, such as the phrase “so help me God”. The lexical semantics of legal-social framing (how constituents perceive the constitution and the law) between de jure and de facto secular constitutions remain debated by legal scholars and political scientists.[31] To account for both arguments, my research includes the two categorizations— de jure and de facto secularity—as independent variables in the data collection process. 

Cultural Context as a Normative Metric

Sociologist Edward Hall’s pioneering work in proxemics—the study of human perceptions of space—and extension transference—the reification process of social phenomena—serve as foundational texts in national semiotics.[32] Proxemics assists semiotic researchers in determining the factors of constructed spatial organizations—in this case, symbols assisting the creation of national communities.[33] Although my research aim shifted from a post hoc study on anthem semiotics to investigating the a posteriori state characteristics of anthem adoption, Hall’s work is still valuable to understand as an independent variable due to its relationship to political symbols.[34]

In his books Beyond Culture and Silent Language, Hall operationalizes new demographic comparisons after measuring the nonverbal and implicit communication required to achieve cultural competency, creating the paradigm of high- and low-context cultures.[35] Published in 1959, Silent Language follows the trend in American linguistics away from structuralism and towards nativism, critiquing the status quo tabula rasa understanding of linguistics.[36] Both Hall and Kittler et al. build on Chomsky’s understanding of biolinguistics by referring to culture as an independent extension of humanity, thereby laying foundation in the semiotic community for the high and low culture dichotomy.[37] My research incorporates this work through Hofstede’s operationalization of cultural context in a study of political organization, which ultimately argues that the nonverbal context of a culture is a significant factor in a person’s perception of individualism and collectivism—and thereby affects their opinion of communal political symbols.[38]


I use the following hypotheses to test my explanatory variables of fascism and authoritarianism:

HA: A country with a history of a significant fascist party will take longer to adopt a national anthem than a country without such a history.

HBA country with a history of authoritarianism will adopt an anthem sooner than a country without a history of authoritarianism.

Research Design

To investigate the association between regime type and time taken to adopt a national anthem, I select variables supported by causal theories relevant to in situ case studies of specific national anthems. I conduct a multiple-variate linear regression analysis with the constructed data set to determine the statistical relationship that each independent variable shares with the number of years that each case takes to adopt a national anthem, thereby executing a cross-sectional analysis to uncover the varying increasing, decreasing, directional, or magnitudinous outcomes. To best measure the sample of cases, I operationalize my dependent variable by calculating the difference of the year of autonomy and first year of anthem adoption for a country. Four countries’ anthems do not feature official lyrics, all of which I include, due to a lack of literature detailing significant semiotic differences between anthems with and without lyrics [39].

I draw each date of autonomy from the CIA World Factbook and the date of de jure anthem adoption from each case’s accessible legislative history. Then I further verify with the National Anthem Creative Commons Database.[40] I reach a rapprochement with my dependent variable through the test and eventual exclusion of various potential paths of inquiry. First, I exclude minor lyrical modifications made to an anthem—present in states such as Bahrain, Cambodia, Chile and the former Soviet Socialist Republics (SSRs)—and only consider the first year of adoption. Second, I limit countries with two de jure equal anthems to their state anthem—such as Denmark and New Zealand—and exclude anthems that are only used for regal occasions as to match the semiotics discipline-wide rejection of such anthems.[41] Third, I take the most recent date of autonomy, initially as a concern over autonomy dates of SSRs. I make this choice to match the date of anthem adoption to that of autonomy as to ensure multifinality in the dependent variable. 

I operationalize fascist parties in the research using the Princeton Election Consortium’s Fascist Movement Project coupled with Paxton’s Five Stages of Fascism.[42] Ideological beliefs of fascist parties often differ from one another based on historical context, necessitating a qualification of data to find consensus on fascist parties.[43] For example, the anti-capitalist rhetoric of the Mobutu regime in Zaire serve contrasting to the Partito Nazionale Fascistacorporatist tendencies. To formulate the metadata for fascist parties, I use the Donno and Lijphart election databases and analyze each case’s electoral history to best ensure the veracity of the final data set.[44] Using these datasets, I mark fascist cases as countries with a history of a domestic fascist party holding a significant share of the government. Per Roberts’ argument that the Nazi party’s first meaningful policymaking came in 1930 with a 18.25% vote share of the Reichstag, I code any general election with a fascist party acquiring greater than 15% of the vote share as significant.[45]

Nevertheless, using fascism as the primary explanatory variable would limit the scope of the project, given its relatively recent emergence. Thus, I adopt authoritarianism as a robustness check and to functionally serve as an independent variable, following the methodological recommendations of Biderman and Batista.[46] Considering the political discourse of authoritarianism, there is a difficulty in operationalizing authoritarianism when using multiple primary source data sets.[47] To counter this, I use Geddes’ history of authoritarianism data to find the regime history for every case.[48]

Because Hall’s scale of cultural context exists as an ordinally scaled variable compared to the other independent dummy variables, I code cultural context as a categorical set of dummy variables in STATA, a statistical analysis software. This methodological decision produces twelve different binary outcomes instead of a single sixteen-point scale. This coding allows me to examine the relationship of each level compared to the difference between state adoption of national anthems and autonomy. To account for statistical significance, I adopt a preliminary stepwise backwards elimination test as a method to determine the explanatory strength of statistically significant variables. I use Halverson’s metrics to operationalize high- and low-context cultures as a sixteen-point scale, methodologically applied to each case using Hall’s extension transferential analysis.[49] Halverson designs the scale as a positive relationship with the level of cultural context, an increase along the scale indicating a greater reliance on nonverbal communication. While this set is comprehensive compared to the other (primarily regional analytical) studies conducted using Hall’s metric, Halverson leaves the South Pacific and Central Asia uncoded, which are thereby left out of the data set for this variable. 

As addressed in the previous section, I operationalize constitutional secularity in two groups: de jure and de facto secular. Because secularity is a negative ideal contingent on the absence of religiosity, I adopt the inverse of necessary conditions for religious constitutions. Therefore, de facto secular constitutions feature references to theist systems, but cannot give special legal status to religious organizations; conversely, de jure secular constitutions cannot feature any references to religious organization or religious values. 

I construct my research design with the purpose of exploring causal mechanisms that lead states to adopt anthems faster or slower compared to the arithmetic mean. In addition, I determine that placing a greater emphasis on statistical significance through a methodological commitment to p-values was less sufficient of an analysis than a procedural discussion of the resulting confidence intervals. I base this methodological decision on the findings of Wasserstein et al. and Amrhein, et al., who advise instead to use a creative and pragmatic analysis of confidence intervals when the large number of independent variables in a cross-sectional analysis limits the utility of P-values.[50] Despite there being a large range in the difference between state adoption of national anthems to state autonomy, bucketing outcomes is unnecessary as to maintain each individual value and to not exacerbate existing differences through the use of subjective logarithmic or linear buckets.[51] The cross-sectional nature of the data set and implausibility of reducing anthems to a context-free researchable vacuum limits the implications of the variate r-squared values, as there is a high likelihood of including confounding variables. Wherry’s empirical findings of the impracticality of the r-squared value while using many independent variables in a cross-sectional semiotic study additionally guides.[52] Thus, analyses of the variables’ confidence intervals, coefficient directions, and magnitudes are used to determine the validity of my hypotheses.

Some neo-positivist scholars deal with concerns of multicollinearity with an automated pairwise backwards elimination of collinear variables.[53] However, as explained by Flom and Cassell and Breaux, there are genuine doubts of the efficacy of computational statistical programs in this effort.[54] Breaux finds that computational programs designed to automatically eliminate insignificant variables have a tendency to unintentionally remove methodologically significant variables that the researcher had previously re-coded as categorical dummy variables. The 16-level ordinally scaled variable of cultural context fits this criterion, bringing concerns of hidden point density to fruition and leading me to discount backwards elimination or any other pairwise selection method. To test for potential multicollinearity, I conduct a Variance Inflation Factor (VIF) test. I then convert the VIF score into tolerance, a metric more commonly used to test in the social scientific community.[55] Mirroring the discipline standard, I consider any tolerance level below 0.1 collinear.[56] I also conduct a pairwise scatterplot test on the cultural context and fascist variables to triangulate the results found from the VIF test for tolerance (Appendix A). Additionally, I construct a pairwise scatterplot test on the metadata; however, these findings were insignificant due to the challenges of analyzing such a large set, as addressed by Carr et al.[57] My operationalization quells concerns of multicollinearity between fascism and authoritarianism, as some significant fascist parties never came to power in their country.

Effects of Variables on Anthem Adoption

Following my compilation of research methods, I begin with a VIF test as to test for multicollinearity. I calculate VIF by (where ­­­­­I draw R2 from the statistic in question), then convert into tolerance, as discussed in the previous section. In my test for multicollinearity, I adopt the standard .1 1/VIF threshold in the VIF test displayed in decreasing order in Table 1.

Table 1: Variance Inflation Factor Test for Multicollinearity

Context 1216.060.062
Context 138.610.116
Context 105.820.172
Context 85.120.195
Context 143.920.255
Context 73.790.264
Context 113.730.268
Context 43.150.318
Context 92.960.338
Context 62.420.413
Context 52.160.463
Context 32.130.469
Context 151.790.557
Context 21.730.579
Legal Secularity1.520.658
Context 161.350.739
Mean VIF3.760.266

The VIF test in Table 2 presents the 12th cultural context level as multicollinear. This outcome is not surprising in the methodological analysis, however, as the 12th level of cultural context controls one-third of the total number of cases, increasing the collinearity with other variables (Appendix A). Based on this test, I do not consider outcomes from the 12th cultural context level because it sits below the established tolerance threshold of .1.

To triangulate the findings of the VIF multicollinearity test regarding each level of cultural context and related outcome, I also conduct a pairwise scatterplot test on the two variables. This test, otherwise known as a scatterplot matrix, serves as an additional visual test of multicollinearity to corroborate where the correlation between the two variables lie (Figure 1). Although the full data set is not independently sufficient to test multicollinearity on its own because of the challenges of hidden point density, it can serve as a visual supplement to the VIF test and the Cultural Context-Difference matrix (Appendix A).

Figure 1: Pairwise Scatterplot Between Difference and Cultural Context

Figure 1 matches the variance inflation test seen in Table 1only displaying high levels of multicollinearity with the 12th and 13th levels of cultural context. Testing the level of cultural context as an ordinally scaled variable instead of separate categorical variables is impossible due to the nonlinear and nonuniform nature of the cultural context data. In other words, the lack of a visible trend from context level 1 to level 16 indicates a variable-wide necessity of conversion to categorical dummy variables to further analyze each level, as displayed in Table 2

Table 2: Linear Regression with Cultural Context Coded as a Categorical Variable

Difference between anthem adoption and autonomy  Coefficient  P > |t|  95% Confidence Interval 
Cultural Context 2-90.0460.345-278.04397.950
Cultural Context 3-102.7780.237-273.81468.258
Cultural Context 4-73.9100.326-222.17974.358
Cultural Context 5-162.7350.064-334.8169.346
Cultural Context 6-100.6630.211-259.08757.760
Cultural Context 7-156.3930.031-298.220-14.566
Cultural Context 8-156.2450.025-292.559-19.932
Cultural Context 9-159.2250.030-303.072-15.377
Cultural Context 10-151.0440.029-286.381-15.708
Cultural Context 11-144.4250.059-294.3085.458
Cultural Context 12-141.0530.030-268.448-13.658
Cultural Context 13-113.7780.086-243.77616.219
Cultural Context 14-15.2790.825-151.743121.186
Cultural Context 15-167.4710.086-359.06524.122
Cultural Context 16-98.6890.407-333.311135.932
De facto Secularity26.6290.205-14.68367.941
De jure Secularity-7.6940.716-49.39434.006

Note: I omit Cultural Context to interpret the coefficients on all the other indicators as relative to the first level, I also omit standard error and t-score to maintain brevity and due to the lack of applicability of the metrics (Appendix B).   

The testable indices in the regression are coefficient, P-value, and confidence interval. The second column in Table 2 establishes the coefficients for each variable, which are used to determine the number of years that a country is expected to take before adopting an anthem compared to the mean (labeled as “Constant”) or Context Level 1 (in the measurements of cultural context). When comparing the variables to the difference in years of adoption and autonomy of each case, any of the variables existing as independently present lead to a state adopting an anthem faster than average—except for de facto secularity.  

I find split results between Hypotheses A examining fascism and Hypothesis B examining authoritarianism. As addressed in the research design, I put little procedural weight on P-values; the findings of fascism and authoritarianism still warrant analysis despite a lack of statistical significance due to the cross-sectional nature of the model and existence of many independent variables. The regression shows that countries with a history of fascism are expected to adopt anthems 16.782 years faster than average. According to this test, Doğan’s theory about fascism’s interaction with national anthems and my Hypothesis A fail to reject the null hypothesis.  However, the confidence interval upper limit of 43.768 indicates that some countries with a history of fascism do take longer than average to adopt anthems, as seen with the Italian Giovenezza in Doğan’s case-based observations.  

Countries with a history of authoritarianism are likely to adopt anthems 15.539 years faster than the average, confirming Hypothesis B. However, similar to the outcome of the fascist variable, the confidence interval lower end of -66.887 and upper end of 37.810 indicates lack of significance for this variable. A noteworthy supplemental finding is the calculated difference between the fascist and authoritarian outcomes, wherein a state with a history of fascism is expected to adopt an anthem 2.243 years sooner than a state with a history of authoritarianism: resulting in the inverse direction of my hypotheses. The lack of explanatory power between my hypotheses could indicate that systems of government are not effective determinants of national anthem adoption.

As discussed in the research design, I do not analyze the 12th level of cultural context. The other levels of cultural context show negative coefficients, indicating that each level is likely to adopt anthems faster than the first level of cultural context. The first level of cultural context includes three countries, one of which (Switzerland) took 433 years to adopt a national anthem. This information shifted the data in comparisons between levels of cultural context, leading to the negative heading of each coefficient. Nonetheless, I can still analyze the coefficients in relation to one another, with the results of levels 7-10 and level 14 worth analyzing because their confidence intervals exist measurably as lower than zero. The coefficient of level 14 is significantly greater than the rest, indicating that a country at this level is likely to adopt an anthem much faster than one of the other levels. Cultural context levels 7-10 are lower than the rest and are proximate to one another; countries operating at these levels are likely to take longer to adopt anthems. Although there is no overall trend in the scale, the fact that these four levels exist so close to one another at one end of the scale is demonstrative of an additional factor. Cultural context levels 7-10 are made up of Britain and it’s former colonies, as well as Latin American, Slavic, and Benelux countries; this displays a unique, unanticipated trend deserving of further exploration.

The variable of constitutional secularity presented interesting results. De facto and de jure secularity displayed oppositional outcomes in relation to the number of years taken to anthem adoption. Countries that give special constitutional status to a specific religion are likely to take, on average, 26.629 years longer to adopt an anthem than secular states, whereas countries that are constitutionally secular are likely to adopt an anthem 7.694 years sooner than secular states. Both measures of secularity include a small confidence interval surrounding zero. De facto secular constitutions were most commonly adopted in (or exported by) the period following the Enlightenment and introduction of the nation state, which also coincides with the creation of national anthems: visible here as Germany and the United Kingdom. De jure secular constitutions, such as in the cases of Oman and Thailand, were often written with a contemptuous attitude towards Enlightenment secularism, alongside an active rejection of Weberian norms such as the national anthem. Thus, the results of both de jure and de facto constitutions can be considered unsurprising.

Reflexivity and Shortcomings

Because I root my research in semiotics, it is important to consider what areas of indexical semiotics differ from my analysis. This requires acknowledging elements of reflexivity most often reserved for research outside of a neopositivist epistemology. First, a discipline-specific shortcoming: the adoption of national anthems is a recent phenomenon. Argentina adopted the first continuous national anthem in 1816, 546 years after the founding of the first continuous state. Although the pieces themselves were often composed centuries before, national anthems were first adopted out of the early 19th century nationalist wave seen in South America.[58] Therefore, in regard to time taken to adopt a national anthem, the data may be skewed towards countries with recent autonomy, such as former colonies. 

There were two shortcomings with the operationalization of fascism and authoritarianism as variables. First, the 15% vote share for fascist countries is rooted in a subjective understanding of political influence. Dumasy argues that fascist parties are often far more influential than their specific vote share indicates; if true, fascist parties may command influence long before they attain this 15% vote share.[59] In my data, however, there are no cases of fascist parties with proximate enough electoral shares to 15% to consider including. Second, the quality of democracy in each fascist state is also relevant. Gooch found that under Italian fascism, political assassinations and overt or covert voter suppression ran rampant, limiting the validity of electoral data acting as a determinant of ideological party support.[60] This is additionally true for authoritarian regimes, though less universal than fascist states.[61] Fascist parties intertwining themselves with other aspects of social life outside of government counter Gooch’s implications for data collection in this project; many parties claim to ideologically disagree while supporting the regime. The unpredictable nature of popular support regarding authoritarian regimes makes controlling for degrees of voter suppression difficult outside of the adoption of a large, well-represented data set. 

To overcome a sampling challenge in how I code fascism (as many states adopted anthems before fascism became a plausible political project), I use founding dates for cases instead of previous dates of autonomy, such as the founding of Zaire. Besides the temporal exportation of national anthems, many anthems have compositionally similar characteristics and are almost universally built out of a Weberian rational-legal model.[62] In anthems that do not follow these characteristics, I recognize that the measurements of state metrics are not entirely explicative for adoption as the justification for anthem adoption may not follow the Weberian model this research attempts to understand.[63]

While each state in the data eventually legally adopted their anthems, potentially reflecting a comparable understanding of national symbols, there is an amount of normative tension surrounding understandings of what national anthems represent. There is not an effective way to proscribe an anthem value to each case, as the state may choose the date of adoption for other reasons if the embodiment of the national anthem differs from that of the Weberian model. In addition, my position within a country that quasi-organically developed within the Weberian rational-legal model may have influenced my choices and theoretical understandings of potential variables, which is in stark contrast to countries where this mode of social organization was violently forced upon. Some argue that the existence of the Weberian model is necessary for anthems to perform ritual social functions (such as membership requirements to the United Nations), to induce childhood socialization into a nation (either with the state or outside of it), to display sentiments of national solidarity, and to inspire enthusiasm in a group.[64] Some believe that this observation is not necessarily inherent to anthems themselves, but rather is a reflection of their environment. In her phenomenological study into the rapid understanding of the Japanese nation, Adal asserts that mimesis, or the interaction between nations, is often “inflected with inequalities of power.”[65] She further argues that national anthems are one of the most observable examples of power dynamics that play a role in mimesis, especially in the reconstructed dialectic between East and West. Further research connecting the Edward Said-brand of construction of the East using anthems would be interesting.[66] Because I did not consult anyone outside of comparable positionalities to myself, my variable selection may display preconceived notions of my socialization.


This paper is a foray into the synthesis of national anthem adoption and adoption semiotics, considering the strength of anthems in conveying policy preferences and socializing citizens. As separatist movements compete with concentrations of power across the world, both right- and left-wing groups continue to find anthems as a useful tool for collective cohesion, thus displaying their continued relevance.[67] Although I only applied this research to the state adoption of anthems, the source of anthem meaning is often in an ethnic or sectarian identity that precedes the state.[68] Thus, further research should be considered on anthem semiotics through the continuation of an expansion or modification of the variables tested in my analysis. This research attempts to conceptualize differences between the rationales of states choosing to adopt national anthems. Doğan’s functionalist theory of fascism was not displayed in the data. States with a history of fascism, Hypothesis A, took longer to adopt anthems on average than states without (though the confidence interval displayed a wide range). 

The results of the authoritarian variable failed to reject Hypothesis B, finding that states with an authoritarian past adopt anthems sooner than those without; though there is a range of cases adopting both faster and slower than non-authoritarian regimes. The outcomes are promising for further research within the anthem semiotics community as both fascism and authoritarianism show heterogeneous variation between their cases and the mean. This could result in a future disregard for these forms of government or political organization entirely from further inquiry. On average, countries with de facto constitutional religiosity are much more likely to take longer in adopting a national anthem than countries with an element of legal religiosity. Regarding cultural context, there was no cross-variate trend on the ordinal measurement, although levels 2-16 of cultural context were significantly more likely to adopt anthems faster than the first level. 

This study offers a foundation for further research into the politics and semiotics of national anthems. I specifically encourage research into how other variables, such as economic indices or violent struggles for autonomy, interplay with anthem adoption. Until the nation-state ceases to serve as the foundational unit of international relations, national symbols such as anthems will continue to perform their role as instruments to forcibly aggregate the two moieties. Studying the development of national anthems across time and space can teach us about the early foundations of the nation-state and allow us to make inferences into the origins and future outlook of the imagined political community. 

Appendix A

Appendix B

VariablesStd. Errort
Context 295.139-0.95
Context 386.556-1.19
Context 475.034-0.99
Context 587.085-1.87
Context 680.173-1.26
Context 771.774-2.18
Context 868.984-2.26
Context 972.797-2.19
Context 1068.490-2.21
Context 1175.851-1.90
Context 1264.471-2.19
Context 1365.788-1.73
Context 1469.060-0.22
Context 1596.960-1.73
Context 16118.735-0.83
Legal Secularity21.103-0.36


[1] F. Thomas Trotter, “On Being Alive to the Arts and Religion: Music – Religion Online,” Religion Online (blog), 1987, https://www.religion-online.org/article/on-being-alive-to-the-arts-and-religion-music/.

[2] Daniel Hummel, “Banal Nationalism, National Anthems, and Peace,” Peace Review 29, no. 2 (April 2017): 225–30, https://doi.org/10.1080/10402659.2017.1308736.

[3] Avi Gilboa and Ehud Bodner, “What Are Your Thoughts When the National Anthem Is Playing? An Empirical Exploration,” Psychology of Music 37, no. 4 (October 2009): 459–84, https://doi.org/10.1177/0305735608097249.

[4] Raja Adal, “Interlude.: Mimesis and Seduction in National Anthems,” in Beauty in the Age of Empire, Japan, Egypt, and the Global History of Aesthetic Education (Columbia University Press, 2019), 105–19, https://doi.org/10.7312/adal19120.11.

[5] Raymond Firth, Symbols: Public and Private, Routledge Revivals (Abingdon, Oxon: Routledge, 1973).

[6] Gilboa and Bodner, “What Are Your Thoughts When the National Anthem Is Playing?”

[7] “The World Flag Database: What’s New,” accessed April 22, 2020, http://www.flags.net/whatsnew.htm.

[8] Toby Perkins, “Toby Perkins MP: It is time to let England’s voice be heard,” PoliticsHome.com, January 27, 2016, https://www.politicshome.com/news/uk/culture/opinion/house-commons/59108/toby-perkins-mp-it-time-let-englands-voice-be-heard.

[9] Tim F. Liao, Gehui Zhang, and Libin Zhang, “Social Foundations of National Anthems: Theorizing for a Better Understanding of the Changing Fate of the National Anthem of China,” Journal for the Theory of Social Behaviour 42, no. 1 (2012): 106–27, https://doi.org/10.1111/j.1468-5914.2011.00480.x.

[10] Adal, “Interlude.”

[11] Donald J. Cunningham, “Beyond Educational Psychology: Steps Toward an Educational Semiotic,” Educational Psychology Review 4, no. 2 (1992): 165–94.

[12] Michael Pressley, Marilyn B. Marks, and Pamela Beard El-Dinary, “Beyond Educational Semiotics: A Response to Cunningham,” Educational Psychology Review 4, no. 2 (1992): 241–52.

[13] Vyach. Vs. Ivanov and Doris Bradbury, “The Science of Semiotics,” New Literary History 9, no. 2 (1978): 199–204, https://doi.org/10.2307/468569.

[14] Maria Isabel Mayo-Harp, “National Anthems and Identities: The Role of National Anthems in the Formation Process of National Identities” (M.A., Canada, Simon Fraser University (Canada), 2001).

[15] Deborah Bradley, “Oh, That Magic Feeling! Multicultural Human Subjectivity, Community, and Fascism’s Footprints,” Philosophy of Music Education Review 17, no. 1 (2009): 56–74; Carlos R. Abril, “Functions of a National Anthem in Society and Education: A Sociocultural Perspective,” Bulletin of the Council for Research in Music Education, no. 172 (2007): 69–87.

[16] Luca de Caprariis, “‘Fascism for Export’? The Rise and Eclipse of the Fasci Italiani All’Estero,” Journal of Contemporary History 35, no. 2 (2000): 151–83.

[17] Derya Karaburun Doğan, “The Role of Music in Fascism Propaganda: The Example of the Mussolini Italy / Faşizm Propagandasında Müziğin Rolü: Mussolini İtalyası Örneği,” Folklor/Edebiyat 25, no. 97 (2019): 229–43, https://doi.org/10.22559/folklor.368.

[18] Derya Karaburun Doğan.

[19] Alexandru-Ovidiu Vintila, “TRAIAN BRAILEANU, GRUPUL DE LA CERNAUTI SI REVISTA ‘ÎNSEMNARI SOCIOLOGICE’ (DEBUTUL UNOR DEZBATERI) – ProQuest,” Revista Romana de Sociologie, 9, 2012, 483–501.

[20] Ali E. Erol, “The Internarrative State” (Ph.D., United States — District of Columbia, Howard University, 2013), https://search.proquest.com/docview/1439925313/abstract/BF8654888B8D4439PQ/1.

[21] Philip Roessler, “THE ENEMY WITHIN: Personal Rule, Coups, and Civil War in Africa,” World Politics 63, no. 2 (2011): 300–346.

[22] William H. Epstein, “Introduction: Biopics and American National Identity—Invented Lives, Imagined Communities,” A/b: Auto/Biography Studies 26, no. 1 (2011): 1–33, https://doi.org/10.1353/abs.2011.0001.

[23] J. Martin Daughtry, “Russia’s New Anthem and the Negotiation of National Identity,” Ethnomusicology 47, no. 1 (2003): 42–67, https://doi.org/10.2307/852511.

[24] Gilboa and Bodner, “What Are Your Thoughts When the National Anthem Is Playing?”

[25] Bradley, “Oh, That Magic Feeling! Multicultural Human Subjectivity, Community, and Fascism’s Footprints.”

[26] Johan Goerg Lughofer, “Cultural Diversity Vs. National-Socialist Positions. Paula Von Preradović: An Istrian and the Author of the Austrian National Anthem,” Acta Histriae 29, no. 3 (2019): 495–512.

[27] Stanley G. Payne, “Fascism and Right Authoritarianism in the Iberian World – The Last Twenty Years,” Journal of Contemporary History21, no. 2 (1986): 163–77; Daughtry, “Russia’s New Anthem and the Negotiation of National Identity.”

[28] Liam D. Murphy, “Calling It Mammon: Instrumentalised Secularity and Religious Futures in Northern Ireland,” Anthropological Journal of European Cultures 20, no. 2 (2011): 114–37.

[29] Katherine Meizel, “A Singing Citizenry: Popular Music and Civil Religion in America,” Journal for the Scientific Study of Religion 45, no. 4 (2006): 497–503.

[30] “Constitute,” accessed May 1, 2020, https://www.constituteproject.org/.

[31] Matthew S. Holland, “‘This Nation, Under God,’” in Bonds of Affection, Civic Charity and the Making of America–Winthrop, Jefferson, and Lincoln (Georgetown University Press, 2007), 200–218, https://www.jstor.org/stable/j.ctt2tt31g.13; Perry A. Zirkel, “Under God?,” The Phi Delta Kappan 84, no. 10 (2003): 796–97.

[32] Edward T. Hall, Beyond Culture, Anchor Books edition. (New York: Anchor Books, 1981).

[33] Nelson González Ortega, “The Formation of Political Traditions and National Symbols in Nineteenth-Century Latin America,” Romance Studies 35, no. 1 (January 2, 2017): 59–72, https://doi.org/10.1080/02639904.2017.1306334.

[34] Geert H. Hofstede, Culture’s Consequences: International Differences in Work-Related Values, Cross-Cultural Research and Methodology Series 5 (Beverly Hills, Calif: Sage Publications, 1980).

[35] Hall, Beyond Culture; Edward Twitchell Hall, The Silent Language (Doubleday, 1959).

[36] Hall, The Silent Language; Noam Chomsky, “Logical Structures in Language,” American Documentation 8, no. 4 (1957): 284–91, https://doi.org/10.1002/asi.5090080406.

[37] Hall, Beyond Culture; Markus G. Kittler, David Rygl, and Alex Mackinnon, “Special Review Article: Beyond Culture or beyond Control? Reviewing the Use of Hall’s High-/Low-Context Concept,” International Journal of Cross Cultural Management 11, no. 1 (2011): 63–82, https://doi.org/10.1177/1470595811398797.

[38] Hofstede, Culture’s Consequences.

[39] Adal, “Interlude.”

[40] David Kendall, “National Anthem Database,” nationalanthems.info, 2019, http://www.nationalanthems.info/.

[41] Piet Strydom, “Philosophies of the Social Sciences,” 2011, 95–125.

[42] Wang Sam, “What Actions Are Shared to All Fascist Movements?,” Princeton Election Consortium (blog), December 21, 2016, http://election.princeton.edu/2016/12/21/what-actions-are-shared-to-all-fascist-movements/; Robert O. Paxton, “The Five Stages of Fascism,” The Journal of Modern History 70, no. 1 (1998): 1–23, https://doi.org/10.1086/235001.

[43] Roessler, “THE ENEMY WITHIN.”

[44] Daniela Donno, “Replication Data for: Elections and Democratization in Authoritarian Regimes,” American Journal of Political Science(Harvard Dataverse, May 26, 2015), https://doi.org/10.7910/DVN/UJYLT0; Gary Cox and Arend Lijphart, “Lijphart Elections Archive: Lijphart Country Catalog: Costa Rica” (Harvard Dataverse, January 19, 2009), https://doi.org/10.7910/DVN/CUYGEY.

[45] David D. Roberts, “Internal Interaction:: Fascists, Conservatives, and the Establishment,” in Fascist Interactions, 1st ed., Proposals for a New Approach to Fascism and Its Era, 1919-1945 (Berghahn Books, 2016), 61–124, https://www.jstor.org/stable/j.ctvpj7hpg.7.

[46] Ciro Biderman and Yuri Camara Batista, “Robustness Check,” Is Progressive Property Tax Progressive? (Lincoln Institute of Land Policy, 2018), JSTOR, https://www.jstor.org/stable/resrep18616.7.

[47] Jos D. Meloen, “Authoritarianism in The Netherlands: Mission Completed? Downward Trends in Authoritarianism in The Netherlands 1970-1992 with an International Comparison of World Data,” Social Thought & Research 22, no. 1/2 (1999): 45–93.

[48] Barbara Geddes, Joseph Wright, and Erica Frantz, How Dictatorships Work – Autocratic Regimes (Cambridge University Press), accessed March 19, 2020, https://sites.psu.edu/dictators/how-dictatorships-work/; Barbara Geddes, Joseph Wright, and Erica Frantz, “Autocratic Breakdown and Regime Transitions: A New Data Set,” Perspectives on Politics 12, no. 2 (June 2014): 313–31, https://doi.org/10.1017/S1537592714000851.

[49] Halverson C.B., Cultural-Context Inventory: The Effects of Culture on Behavior and Work Style, J.W. Pfeiffer, vol. Annual Developing Human Resources (Pfeiffer and Company, San Diego, CA., 1993), https://www2.pacific.edu/sis/culture/pub/context_cultures_high_and_lo.htm; Hall, The Silent Language.

[50] Ronald L. Wasserstein, Allen L. Schirm, and Nicole A. Lazar, “Moving to a World Beyond ‘p < 0.05,’” The American Statistician 73, no. sup1 (March 29, 2019): 1–19, https://doi.org/10.1080/00031305.2019.1583913; Valentin Amrhein, Sander Greenland, and Blake McShane, “Scientists Rise up against Statistical Significance,” Nature 567, no. 7748 (March 2019): 305–7, https://doi.org/10.1038/d41586-019-00857-9.

[51] B. D. Ripley, “Spatial Statistics: Developments 1980-3, Correspondent Paper,” International Statistical Review / Revue Internationale de Statistique 52, no. 2 (1984): 141–50, https://doi.org/10.2307/1403097.

[52] R. J. Wherry, “A New Formula for Predicting the Shrinkage of the Coefficient of Multiple Correlation,” The Annals of Mathematical Statistics2, no. 4 (1931): 440–57.

[53] John D. Morris, Faye B. Morgan, and Wilbert Maynor, “On Selecting the Best Set of Regression Predictors,” The Journal of Experimental Education 48, no. 2 (1979): 100–103.

[54] Peter L Flom and David L. Cassell, “Stopping Stepwise: Why Stepwise and Similar Selection Methods Are Bad, and What You Should Use,” NESUG 2007, 2007, 7; Harold J. Breaux, “A Modification of Efroymson’s Technique for Stepwise Regression Analysis” (Association for Computing Machinery, August 1, 1968), https://doi.org/10.1145/363567.363580.

[55] Edward R. Mansfield and Billy P. Helms, “Detecting Multicollinearity,” The American Statistician 36, no. 3 (1982): 158–60, https://doi.org/10.2307/2683167.

[56] Morris, Morgan, and Maynor, “On Selecting the Best Set of Regression Predictors.”

[57] D. B. Carr et al., “Scatterplot Matrix Techniques for Large N,” Journal of the American Statistical Association 82, no. 398 (1987): 424–36, https://doi.org/10.2307/2289444.

[58] Mayumi Itoh, “Japan’s Neo-Nationalism: The Role of the Hinomaru and Kimigayo Legislation,” Japan Policy Research Institute Working Paper No. 79 (July 2001), http://www.jpri.org/publications/workingpapers/wp79.html.

[59] François Dumasy, “Le Fascisme Est-Il Un ‘Article d’exportation’? Idéologie et Enjeux Sociaux Du Parti National Fasciste En Libye Pendant La Colonisation Italienne,” Revue d’histoire Moderne et Contemporaine (1954-) 55, no. 3 (2008): 85–115.

[60] John Gooch, “Re-Conquest and Suppression: Fascist Italy’s Pacification of Libya and Ethiopia, 1922-39,” Journal of Strategic Studies 28, no. 6 (2005): 1005–32, https://doi.org/10.1080/01402390500441024.

[61] Peter Juviler, “Varieties of Authoritarianism,” in Freedom’s Ordeal, The Struggle for Human Rights and Democracy in Post-Soviet States (University of Pennsylvania Press, 1998), 82–103, https://www.jstor.org/stable/j.ctt3fh7sb.10.

[62] Stanford CCRMA, “Features of National Anthems,” Center for Computer Research in Music and Acoustics, accessed April 20, 2020, https://ccrma.stanford.edu/~linford/features.html.

[63] Donald Eugene Smith, “Religion and Politics in Burma,” in Religion and Politics in Burma (Princeton University Press, 1965), 230–80, https://doi.org/10.2307/j.ctt183pz1j.10; Karen A. Cerulo, “Sociopolitical Control and the Structure of National Symbols: An Empirical Analysis of National Anthems,” Social Forces 68, no. 1 (1989): 76–99, https://doi.org/10.2307/2579221.

[64] Aleksandar Pavkovic and Christopher Kelen, Anthems and the Making of Nation States: Identity and Nationalism in the Balkans(Bloomsbury Publishing, 2015).

[65] Adal, “Interlude.”

[66] Rachel Beckles Willson, Orientalism and Musical Mission: Palestine and the West (Cambridge University Press, 2013).

[67] Anderson Alves de Souza, “Investigating Prescriptions of Positive Behavior in Power-Conso¬lidation National Anthems,” Signótica 29, no. 2 (August 30, 2017): 575–601, https://doi.org/10.5216/sig.v29i2.44102.

[68] Liao, Zhang, and Zhang, “Social Foundations of National Anthems”; Mayo-Harp, “National Anthems and Identities.”

Price Phillips
Price Phillips is a third-year senior at the American University School of International Service (SIS). He is finishing a degree in International Studies with thematic emphases on Peace and Conflict Studies and Comparative Governance. Outside of SIS, Price worked as the Chief of Staff at the American University International Relations Society and as an oversight intern at the Senate Committee on Foreign Relations.