Abstract
This comparative study examines the interplay of religious messaging and disinformation in the election campaign material of Jair Bolsonaro and Recep Tayyip Erdogan for the 2022 Brazilian and 2023 Turkish presidential elections. We employ a mixed-methods approach, combining computational keyword filtering and content analysis with qualitative discourse analysis and applied to a corpus of 10,519 posts across seven social media platforms. The analysis informs two key findings. First, in both Bolsonaro and Erdogan’s presidential campaigns, religious rhetoric and symbolism is used to bolster personal authority and in-group support to consolidate the idea of (presiding over) a majority bloc. Second, while disinformation interfaces with religious messaging in both presidential campaigns, we find that religion is overwhelmingly instrumentalized as a vector for political disinformation—thus, we did not identify any significant presence of “religious disinformation,” classically understood. Ultimately, our article creates space for further comparative and region-specific scholarship to explore how religion, politics, and disinformation intersect with authoritarian populism in modern elections and beyond, while also contributing to the conceptual refinement of “religious disinformation.”
Introduction
In recent years, we have seen an explosion of interest in how disinformation interfaces with democratic outcomes, ranging from studies of disinformation campaign anatomies (Keller et al. 2020) and external election interference (Lysenko and Brooks 2018) to the role of disinformation in creating alternative ontologies that catalyze political demonstration (Prochaska et al. 2023). Despite these advances, we have much to learn about how religious disinformation is strategically leveraged by election candidates, especially incumbents who cannot so easily present themselves as anti-establishment avatars for altruistic rage. Contemporary elections provide an opportunity to evaluate these dynamics in modern political fora.
This article detects and compares the instrumentalization of religious discourse in two countries from contrasting geographical regions: Brazil and Turkey. Specifically, it looks at the interplay between religion and disinformation as contained in the cross-platform campaign discourse of incumbent Presidents Jair Bolsonaro and Recep Tayyip Erdogan during the 2022 Brazilian Presidential Elections and 2023 Turkish Presidential Elections (for detailed information, please see the Supplemental Information File: Appendix A). The article begins with a review of the literature on the function(s) of religion in campaign strategies and the role played by disinformation therein, before narrowing in on the category of “religious disinformation” and the core assumptions that underpin it. Thereafter, we outline our methodological approach, which combines computational analysis with quantitative and qualitative discourse analysis, applied to 10,519 posts across seven social media platforms, which allow us to meaningfully reflect on partisanship, identity politics, and local information contexts (Oehmer-Pedrazzi et al. 2023). Having outlined the data gathering and filtering processes, along with our statistical analysis, the article presents a discourse analysis of the filtered corpus, which is guided by Ernesto Laclau’s theoretical framework on populism being a logic of articulation—that is, we identify how religion and disinformation interact within a general framework of populist articulation without positioning the article as a study that prioritizes populism as its referent object (de la Torre 2019; de Vreese et al. 2018; Ernst et al. 2019; Marzouki 2022). The discourse analysis is presented in a narrative-analytic style, which situates the campaign rhetoric its politico-religious context through ample exposition of primary discourse, while attempting to ensure readability (see Fitzgerald 2014).
Our data and analysis inform two key findings. First, in the cases of both Bolsonaro and Erdogan, religious rhetoric and symbolism are clearly used to consolidate personal authority, in-group support, and the impression of presiding over a majority—but (a) religious discourse is not systematically deployed to attack the political opposition in either case; and (b) Bolsonaro (perhaps surprisingly) employs a more concentrated religious vocabulary than Erdogan in order to shape his identity as being politically distinct from his opponents as a messianic leader. Second, disinformation is systematically leveraged against political opposition in both cases, but we find that such attacks are couched in grievances that do not meet the criteria for “religious disinformation,” as typically presented in the literature. To that extent, our analysis problematizes core assumptions about what constitutes “religious disinformation” and how such processes unfold. Ultimately, our article creates space for further comparative and region-specific scholarship to explore how religion, politics, and disinformation intersect in a dynamic communicative environment dominated by social media.
Literature Review
The literature addressing disinformation tends to operate under the assumption of an enforceable boundary and a readily policeable epistemic border between false and correct information. The large body of knowledge on disinformation conceptualizes the phenomenon in various ways, with a broad vocabulary to define falsehoods and deception (Tandoc et al. 2018). However, it does not often offer a model to easily guide the application of the concept in practical, “real” situations. Highlighting the reliance on “expert consensus” and “body of evidence or facts” to define disinformation may be viable in domains of health or science, but it is much more difficult to use for ideological topics and social beliefs (Vraga and Bode 2020).
Many scholars (Freelon and Wells 2020; Guess and Lyons 2020; Kuklinski et al. 2000; Vraga and Bode 2020) define mis/disinformation as occurring when “people hold inaccurate beliefs and do so confidently” making a distinction between lack of knowledge or ignorance (misinformation) and confident but inaccurate knowledge based on “inaccurate beliefs” (disinformation). However, this common-use definition does not clarify what “inaccurate beliefs” are and fails to adequately capture the normalization of these beliefs in different societies and cultures (Pasek et al. 2015). There is also considerable disagreement among researchers regarding the role of incentives, intentions, type and extent of harm, the degree of falsehood, and the nature of belief in defining disinformation (Berinsky 2023; Guess and Lyons 2020; Roozenbeek et al. 2022). While numerous articles from the so-called “Global North” concentrate on conspiracy theories and the impact of these beliefs on ideological bias (Berinsky 2023), there is limited discussion on the instrumentalization of institutionalized religions for political manipulation and campaigns.
Accordingly, the definitional parameters of religious dis/misinformation are quite broad. For example, Al-Zaman perceives it as misleading information “related to one or more facets of religion, such as religious figures, practices, scriptures and beliefs” (Al-Zaman 2024: 7). Likewise, Haque et al. conceptualize it simply as “misinformation targeting religious sentiments” (Haque et al. 2020: 4). Despite this opacity, we can assert that “religious disinformation” presents faith-based narratives that include misleading guidance, misinterpreted scriptures and/or religious records, or false claims of divine awareness or power (Alimardani and Elswah 2020). Indeed, we know that politicians in Latin America (Tumber and Waisbord 2021), the Middle East, and North Africa (Alimardani and Elswah 2020) often use fear, emotional appeals, populist rhetoric, and the credibility of religious authority to persuade the recipient of these various messages. Their aim is to blur the boundary between God and the leader and, thus, to “divinify” themselves—a strategic endeavor sharpened by McPhetres and Zuckerman’s (2017) observation that “religious people tend to require less evidence whenever a claim is presented in a non-scientific context” (Druckman et al. 2011: 641). Combined with the fact that users tend to react to religious disinformation more emotionally than is the case with nonreligious content (thereby accentuating the essential power of false and misleading claims (Al-Zaman 2021; Soll 2016)), religious dis/misinformation possesses a unique capacity to persuade; especially when targeted at existing communities of believers (Levin 2020).
The sacralization of politics as a calculated, deliberate, and partisan use of faith by political leaders—the “God strategy” as it is called by Domke and Coe (2010)—is widely used in political disinformation campaigns because it works so well. Caramancion (2023) analyzed how users’ religion contributes to their susceptibility to falling prey to misinformation and disinformation attacks. His findings show that Christians were more susceptible to disinformation compared to Atheists, possibly due to the influence of religious beliefs on information processing. Pretus et al. (2023) highlighted the role of political devotion in sharing misinformation on social media, particularly around sacred moral issues, which can be closely tied to spiritual convictions. The authors show that appealing to sacred values in political messages increases the likelihood of sharing disinformation. Echoing McPhetres and Zuckerman’s (2017) and Bronstein et al. (2019) found that “delusion-prone individuals” may be more likely to accept even delusion-irrelevant implausible ideas and increased belief in “fake news” headlines because of their tendency to engage in less analytic and actively open-minded thinking. The authors’ exploratory analyses showed that dogmatic individuals and religious fundamentalists were also more likely to believe “false news.”
Besides a substantial body of knowledge on the psychological and cognitive side of mis/disinformation, there is a lack of empirical analysis concerning religious disinformation as both a resource and a significant manipulative strategy employed by public figures for political gains. Insufficient scholarly research has been devoted to comprehensively exploring the characteristics, intentions, nature of harms, outcomes, and consequences of religious disinformation within the broader context of the contemporary epistemic and democracy crises. Notably, the existent literature that does deal with these approaches is concentrated in the so-called “Global South.” Instead of framing the problem as one of conspiracy theories connected with community-held beliefs like most North-Western research, scholars from the Majority World observe that religion is commonly instrumentalized by populists’ leaders that use religious elements to validate disinformation. For example, religious disinformation has been observed on social media platforms causing fear, confusion, and polluting the Middle East and North Africa (MENA) online sphere since the coronavirus pandemic. Exploring cases of religious clickbait in the form of false hadiths and viral religious advice from religious figures entrenched in the MENA’s political elite, Alimardani and Elswah (2020) discuss how new dynamics for religion in the age of the Internet are contributing to a uniquely regional and religious form of disinformation.
In the absence of a sufficient body of literature that interrogates the role(s) of religion as a key issue for information integrity, narrow case studies of India, Bangladesh, and Russia become essential, empirical lodestars. In the case of India, Al-Zaman (2021) demonstrates the use of target-specific religious groups to promote political ideologies. The use of digital disinformation in Bangladesh has also been linked to communalism and violence against religious minorities (Al-Zaman and Noman 2023). Indeed, Al-Zaman and Noman (2023) identified that, in Bangladesh, most users react to religious disinformation emotionally, encouraging “destructive reactions” like xenophobic expressions and persecution of out-groups. The domestic manipulation of religion for disinformation in Russia is further exemplified again during the Covid-19 pandemic, where reactionary political groups such as the Moscow Patriarchate, its conservative affiliates and fundamentalist monks have thus taken the lead in producing conspiracy theories, highlighting the multiplicity of Orthodox Christian-based perceptions of the world and the complex continuum between the Russian Orthodox Church’s official positions (Laruelle and Grek 2021).
In the case of Turkey, Yilmaz (2023) discusses how religious populism has been instrumentalized to justify the ruling party Justice and Development Party’s (AKP’s) digital authoritarianism. By framing its actions as in keeping with religious values and morality, the AKP aims to secure support, stifle opposition, and position itself as the singular guardian of Islamic principles and the protector of the nation’s devout youth. This strategic manipulation of religion enables the AKP to fortify its populist authoritarianism, consolidate power, and maintain a dominant presence in the digital landscape. The author also highlights the role played by state-controlled religious institutions, traditional media, social media platforms, as well as religious leaders and organizations in shaping public opinion, thereby enabling the government to exercise heightened control over the dissemination of information.
In Latin America, religious disinformation has also been highlighted as a burgeoning problem. However, compared with other regions in the world, relatively little has been written about locally circulating disinformation (Siles et al. 2021). Enhanced understanding(s) of how disinformation and populism operate in the region requires, we argue, adequate assessment of how religion association with polarizing messages and messianic authority and while this constitutes a core contribution of our article, it builds on a small, but significant body of scholarship. For example, the circulation of disinformation in Latin America has been linked to major social and political events in many of these countries, notably presidential elections. Siles et al. (2023) show how presidential candidates like Fabricio Alvarado and Nayib Bukele used a particular style of communication that blended populist elements with religious discourse during the elections in Costa Rica (2018) and El Salvador (2019). Guevara (2020), Siles et al. (2023), and Santini et al. (2021) also demonstrated that, during the presidential elections that took place in 2018 in Brazil and Mexico, social media fueled a polarization of the electorate on topics such as sexual orientation and moral values. Additionally, the combination of disinformation, populist styles of communication, and religious discourse has been a fertile ground for the rise of political/religious candidates and politicians from neo-Pentecostal churches in the region (Siles et al. 2021). The examples of Jair Bolsonaro in Brazil and Fabricio Alvarado in Costa Rica serve as illustrations of how the ascendance of neo-Pentecostal and evangelical churches has introduced nuance and complexity into the political arena. Key in this process is the construction of a discursive premise around religious identity that separates “us” from “them,” based on the distinction between traditional values and new threats, forming in-groups and out-groups. This symbolic violence was crucial not only to the results of these elections but also how they unfolded (Guevara 2020; Siles et al. 2021). Accordingly, the Bible has become a crucial symbol of populist politics in Latin America (Barbosa and Casarões 2022; Ramirez 2014).
Despite clear advances in the literature, most existing studies have focused on a single country or consider different countries in the same region, especially because religion has strong local characteristics and meanings. In contrast, there is a lack of comparative studies between countries in completely different areas of the world and with significantly diverged cultures and historical backgrounds. Likewise, the basis upon which to classify disinformation that orbits/speaks to religious issues and/or religious audiences as constituting “religious disinformation” remains unclear. Continued use of the concept in this fashion risks essentializing all disinformation in religious settings as being “religious disinformation” and possibly expanding the concept’s capacity, rather than refining it. This study’s primary contributions to the literature, therefore, are to address the lack of comparative analysis of religion-facing disinformation between two significantly diverging countries: Brazil and Turkey. In the process, we contribute to a more nuanced understanding of “religious disinformation,” which, helps our fields to unpack the intertwinement of religion, politics, and disinformation as an increasingly definitive component of modern electoral politics.
Methods and Data Collection
Step 1: Data Collection and Filtering
Our case comparison is primarily justified on the similar political proclivities in Erdogan and Bolsonaro’s discourse and action(s), classified as “authoritarian populism” (e.g., Esen and Gumuscu 2023; Mendes Motta and Hauber 2023; Rojas et al. 2019; Yilmaz and Erturk 2021; please see Supplemental Information File: Appendix A for detailed country analysis). Both have been said to fuel widespread anxiety about safeguarding traditional values perceived as threatened by secularizing forces and have been considered figureheads of theocratic democracies (Saiya 2023). Perceived as strongmen, these political leaders have been criticized for their autocratic tendencies, being considered overwhelmingly illiberal and openly authoritarian (Nai and Toros 2020). However, previous research has shown that their policy responses have varied amidst crisis, mainly due to the complex interaction of social, political, and economic pressures and expectations (Bayar et al. 2023), thus indicating the need for understanding how this variation intersects with their reelection objectives as incumbents.
Against this backdrop, we collected and analyzed Jair Bolsonaro and Liberal party (PL), and Recep Erdogan and AKP 1 and social media posts and sponsored content throughout their presidential campaigns on seven different platforms (see Figure 1 for research design). Preceding studies (Tandoc et al. 2019) have confirmed that users perceive social media platforms as communicative integrated environments, with shared and unique affordances. Building upon this premise, we conducted a multiplatform analysis, acknowledging that users routinely navigate across different social media. Our methodological choices have also taken into account the practices and impacts of platforms and communication spaces differ between countries (Valenzuela and Santos 2024), underscoring the value of multiplatform analysis for comparative studies.
Research design.
The data timeframe was based on the chronology of the electoral campaigns: between August 15 to October 30, 2022, for Brazil and between February 15 to May 28, 2023, for Turkey from Facebook, Instagram, Meta Ads, Telegram, TikTok, Twitter, YouTube, and YouTube TV (only for Brazil). Once official channels and profiles for both candidates and their parties were mapped, we employed different tools for data collection (for details about data collection please see Supplemental Information File: Appendix B).
In order to identify religious disinformation in the incumbents’ campaigns, we developed a keyword list to query the collected social media posts based on embedding technologies, a recommended component of disinformation detection processes (Oehmer-Pedrazzi et al. 2023). Based on Bolsonaro and Erdogan’s speeches in religious events and ceremonies, we established a religious vocabulary (see Supplemental Information File: Appendix B for more details on the procedures for data filtering), which allowed us to identify 549 Bolsonaro and 676 Erdogan publications that mentioned at least one religious keyword. These posts were later subject to qualitative content analysis, in which native researchers systematically accounted for the prevalence of religious content, attacks against outgroups, nationalist rhetoric, and disinformation (Figure 2). Coders verified if the posts were indeed introducing religious claims, promoting religious identities, targeting specific religious communities, or portraying the candidates as majoritarian-faith leaders (Saiya 2023). Additionally, they coded the content as disinformation when they identified intentional use of false, decontextualized or distorted information to manipulate public opinion, discredit enemies, and/or introduce sensationalist views (Fallis 2015; Wardle and Derakhshan 2017). These analytical categories were assessed separately, and the prevalence of disinformation was identified only among the religious posts.
Bar graph visualizing the collected posts with religious and nonreligious content. 2
Step 2: Statistical Analysis
The analysis was conducted using R Studio Version 2024.04.1+748 R Studio Team (2024), with the following packages: dplyr for data manipulation, ggplot2 for data visualization, MASS for fitting negative binomial regression models, and stats for conducting proportion tests and calculating chi-square statistics (the R code for data analysis is available in the GitHub repository). We conducted two statistical analyses to (a) proportions test to determine whether religious posts had a higher propensity to appear on each platform compared to nonreligious posts and (b) negative binomial regression to understand the relationship between engagement and content type (religious or nonreligious) with each platform. The hypotheses for the first analysis are
H0: There is no difference in the proportion of religious posts compared to nonreligious posts on the platform.
H1: There is a difference in the proportion of religious posts compared to nonreligious posts on the platform.
Chi-square tests were performed to evaluate these hypotheses.
For the second analysis, we analyzed the engagement distribution for the posts using histograms (Supplemental Information File: Appendix C, Figures C1 and C2) to visualize the skewness and distribution patterns. The variance mean ratio (VMR) was calculated to assess the data’s dispersion, indicating that Negative Binomial Models were appropriate due to the high VMR (>1) (Supplemental Information File: Appendix C). In order to control for time, we tested two models (see Supplemental Information File: Appendix Tables C1 and C2).
Step 3: Discourse Analysis
As part of our content analysis, we extracted a “contextual repository” of media content, which comprised “packages of discourse” that demonstrate the use of populist rhetoric, combined with religious arguments and nationalist discourse to endorse disinformation narratives across different platforms. These packages are contextual snapshots of campaign content, comprising image, video, text, and audio. This repository provides a transparent “dataset” of content subject to in-depth discourse analysis (included in Supplemental Information File: Appendix D) and serves three functions. First, it provides immediately accessible context for readers to evaluate the authors’ interpretation(s) of the discourse—context that is often obscured in traditional discourse analyses that are restricted by scope and word count (Fitzgerald 2014: 68–77); second, it provided for added reflexivity between the authors, helping to (further) amalgamate our diverse backgrounds and expertise toward shared research outcomes; third, the repository can be leveraged for future studies which seek to further unpack the intersections of religion and disinformation in the context of election campaigns.
Concurring with literature that classifies Bolsonaro (Mendes Motta and Hauber 2023; Rojas et al. 2019) and Erdogan (Esen and Gumuscu 2023; Yilmaz and Erturk 2021) as exemplars of “authoritarian populism,” our qualitative discourse analysis is guided by the work of Ernesto Laclau, which treats populism not as an essential characteristic, but as a logic of articulation:
We only have populism if there is a series of politico-discursive practices constructing a popular subject, and the precondition of the emergence of such a subject is [. . .] the building up of an internal frontier dividing the social space into two camps. But the logic of that division is dictated [. . .] by the creation of an equivalential chain between a series of social demands in which the equivalential moment prevails over the differential of the demands. [Emphasis added] (Laclau 2005: 44)
Though a deep exploration of Laclau’s theory of populism is beyond the scope of this article—but having been exhaustively explored in a previous study (Fitzgerald, 2014)—applying this framework determines two outcomes. First, we do not intend to evaluate the degree to which Bolsonaro or Erdogan are “populist” leaders. Rather, we see these leaders and their associated movements as having “already defined themselves as populist by their common reference to the people” (Westlind 1996: 60). Second, and as a result, populism is not our primary object of analysis. Rather, we engage it as “a tool of discourse analysis” (Stavrakakis 2005: 229)—that is, as a hermeneutic device to tease out key discursive linkages (and silences) that arise in two sets of campaign discourse inalienably focused on “the people” and especially focused on how equivalential chains are constructed through a series of political demands that “simplifies political struggle into an antagonism between ‘us’ and ‘them’, ‘good’ and ‘evil’” (Stavrakakis 2005: 234).
Results
Religious Posts and Platforms—Bolsonaro
The proportions test for Bolsonaro’s social media content and the presence of religious content for each platform demonstrate that the platforms with significant difference in the proportion of religious posts compared to nonreligious posts are Telegram, Twitter, YouTube, and YouTube TV (Table 1 and Figure 3 for visualization of significant results).
Proportion Test of the Association Between [Platform] and Religious Posts: Bolsonaro.
Visualization of the significant results of the proportions test for Bolsonaro.
Religious Posts and Platforms—Erdogan
We ran a proportions test for Erdogan’s social media content and the presence of religious posts across all platforms. Results confirm that only Meta Ads demonstrated a significant difference in the proportion of religious posts compared to nonreligious posts (Table 2).
Proportion Test of the Association Between [Platform] and Religious Post: Erdogan.
The number of posts in Meta Ads was very small (N), likely causing the observed significant difference. However, due to the small sample size, we will not report these results.
Engagement, Religious Posts and Platform—Bolsonaro
The results of the negative binomial regression show significant platform effects on engagement (see Figure 4 and Supplemental Information File: Appendix Table C-1). YouTube (incidence rate ratios [IRR] = 1.911), TikTok (IRR = 9.897), Telegram (IRR = 1.507), Meta Ads (IRR = 1.509), and Instagram (IRR = 2.935) positively influence engagement significantly (p < .001). Conversely, Twitter (IRR = 0.144), and YouTube TV (IRR = 0.072) show significant negative effects (p < 0.001). The interaction between religious content and platform engagement reveals a significantly higher engagement on YouTube TV (IRR = 2.494) for religious content, while other platforms show nonsignificant interactions. The overall impact of religious content alone is nonsignificant.
Bolsonaro incidence rate ratios with 95% confidence intervals for the effect of platform and content type on engagement. Blue markers represent a significant positive effect, black markers indicate a significant negative effect, and gray markers show nonsignificant effects. All confidence intervals are at the 95% level.
Engagement, Religious Posts, and Platform—Erdogan
The results from the negative binomial regression model for Erdogan’s social media accounts indicate significant platform effects on engagement (for base model see Supplemental Information File: Appendix Table C-2). YouTube (IRR = 3.408), TikTok (IRR = 9.846), and Instagram (IRR = 2.369) positively influence engagement significantly (p < .001). Conversely, Twitter (IRR = 0.447) shows a significant negative effect (p < .001). The interaction between religious content and platform engagement reveals significantly lower engagement on YouTube (IRR = 0.018), TikTok (IRR = 0.054), Twitter (IRR = 0.328), Instagram (IRR = 0.177), and Telegram (IRR = 0.212) for religious content. Conversely, the overall impact of religious content alone is significant and positive (IRR = 7.539; Figure 5).
Erdogan incidence rate ratios with 95% confidence intervals for the effect of platform and content type on engagement. Blue markers represent a significant positive effect, black markers indicate a significant negative effect, and gray markers show nonsignificant effects. All confidence intervals are at the 95% level.
Discourse Analysis
2022 Brazilian Presidential Elections
In a campaign video uploaded to YouTube on September 10, 2022, set against a background of swelling crowds of Bolsonaro supporters and suspenseful music, the essence of Bolsonaro’s religious framing of political demands is captured in a carefully constructed soundbite:
Today you have a President who believes in God. A government that defends the family. We are a majority Christian homeland, which does not want the liberalization of drugs, which does not want the legalization of abortion, which does not admit gender ideology, and a president who owes loyalty to his people.
3
With almost perfect populist symmetry, Bolsonaro wraps his authority at once (and immediately) in God—but simultaneously in the loyal faith of the people. This dialectical authority is strengthened (and indeed made possible) by a series of familiar oppositional political demands that constitute the core of Bolsonaro’s campaign—reminding the electorate that Bolsonaro’s identity is predicated, always, as being at once with God and “against PT”—a strategy that helped Bolsonaro capture the Presidency in 2018 (see Supplemental Information File: Appendix D). This structure is clearly replicated, for example, in a campaign Instagram post on October 27, 2022 (two days ahead of the crucial second round vote; Figure 6):
The PL Party is (left side) for Family, Christian Values, Reduction of the Minimum Age for Incarceration, Economic Liberalization, Patriotism. Against (right side): Gender Ideology, Abortion, Media Censorship, Attacks on Religion, the Workers Party and Communism. 4
Re-activating long-standing, secular, anti-Workers Party discourses that variously associate the party with crime, corruption, lawlessness, social degradation, and communism (see Santini et al., 2021), Bolsonaro can inject his (relatively new-found) religiosity into an anti-Workers Party chain of equivalence that already resonates with a huge proportion of the electorate: a discursive move that effectively maintains his identity as an outsider, even as he is positioned as the incumbent.
Indeed, owing to a lack of declared religious conviction over the course of his political career, Bolsonaro (unlike Erdogan) cannot derive his “religious authority” from precedent. Instead, his faith is characterized as a revelation. To that end, the “miracle” of having survived a 2018 stabbing at a campaign event in Juiz de Fora features prominently in Bolsonaro’s 2022 election campaign as proof that his mission to lead Brazil to prosperity is one ordained by God:
I want to thank God for my second life that He gave me in Juiz de Fora, not allowing my daughter Laura to be an orphan. I also thank Him for the mission of commanding this country in one of humanity’s most difficult moments. Thank you very much, my God! And if this is Your will, I will be ready to serve another term as President of the Republic.
5
Michele Bolsonaro, a bona fide Evangelical Christian, routinely furthers the religious credentials of her spouse by activating a dense religious vocabulary that nonetheless retains the anti-Workers Party core that is so pivotal to the Bolsonaro campaign. In an August 2022 speech at the site of the 2018 stabbing attempt, for example, the shift to a more millenarian tone is clear—but her framing nonetheless executes the same core appeal to populist demands, while (as always) designating the Workers Party as “thieves” and “enemies of the people”:
With the help of the people, with the help of our good God, our Brazil will emerge victorious. We ask God at this moment for protection from the heavens, for peace, for liberation for those who are deceived. We know that the enemy just wants to steal and destroy . . . Our land is prosperous and blessed . . . This campaign, once again, is a miracle from God, when God performed a miracle in my husband’s life.
6
Regardless of one’s suspicions about the veracity of Jair Bolsonaro’s Evangelical rebirth, it is impossible to classify such positions as being dis/misinformation. Explicit recourse to disinformation is much more apparent in the condemnation of Lula and the Workers Party as being corrupt and communist(/Godless) and hinges, quite substantially, on the notion of “external ties/interference” and associated financial decay. In an official Facebook post on October 4, 2022, for example, Bolsonaro outlines the two “very distinct roads” that lie ahead for the Brazilian people:
In Lula’s path, the government is formed by political appointments, negotiated in the shadows in exchange for support, a prime source of corruption. In Jair Bolsonaro’s path, the government is formed independently, allowing people to work for the country and not for foreign interests. [Emphasis added] In Lula’s path, Brazil is a friend of socialist dictatorships and finances works in Cuba and Venezuela while abandoning those in its own country. Jair Bolsonaro’s path brought Brazil closer to the free and democratic world, signing major agreements and moving towards joining the OECD. [Emphasis added]
7
By tying Lula/the Workers Party to “socialist dictatorships” in Venezuela, Cuba, and Nicaragua, Bolsonaro can effectively leverage “communism” and “socialism” (used interchangeably) as signifiers that foretell the division of the Brazilian people through the importation of social and economic degradation. Indeed, welding the classically pejorative signifiers of “socialism” and “communism” to a modicum of empirical truth forges a conduit through which Bolsonaro can push his most outlandish claims: it is around these categories that the most explicit forms of disinformation appear, including false claims that Lula—mimicking the policies of “socialist dictatorships”—will shut churches, support abortion, and enact a raft of initiatives that would import into a Brazil a clear plan to dilute the Christian institution of the family:
You need to be alert. From today onwards, more than ever, those who love red [supporters of PT] will start wearing green and yellow [supporters of Bolsonaro], those who persecuted and defended closing churches will consider themselves great Christians, those who support and praise socialist dictatorships will say they are defenders of democracy. We have the privilege of not needing to deceive the people about what our values are during this period: we are in favor of the family, the free market and the right to self-defense. We are against drugs and drug trafficking, control of the media and internet, gender ideology and abortion.
8
In the end, Bolsonaro’s religious framing of political demands rests on the clear division of Brazil’s social frontier into two camps (Laclau 2005: 44), with God on his side and the Godless, communist Workers Party on the other. Bolsonaro’s 2022 election campaign, aided by the strategic deployment of disinformation, is one in which religious discourse figures quite heavily—but it does not overwhelm it. Rather, the deployment of religious rhetoric marks an evolution of the political message presented to voters in 2018: a typically populist appeal to the idea of societal breakdown and the consolidation of Bolsonaro as the leader—already chosen by God and the people—to prevent (further) decadence and “return” Brazil to glory:
The Brazilian flag is a symbol of all of us. A few years ago, we saw some people stepping on it, tearing it up and burning it. We rescued this flag, we all rescued our values, our patriotism, our freedom and our faith. There are things that are costly to all of us: abortion, gender ideology, drugs, private property and family. I’m sure that this country called Brazil continues to be the Promised Land . . . Thank you very much, my good God, for this difficult mission that you gave me, through the hands of many gentlemen, to be the head of the federal executive.
9
2023 Turkish Presidential Elections
Although generally seen as a pragmatic “shapeshifter” intent on maintaining power (Genc 2019), Recep Tayyip Erdogan has long presented himself as a defender of Turkey’s Sunni Islamist parties (Genc 2019; Yilmaz 2021), while attaching to his leadership an element of the “divine” (Yilmaz 2021: 7–8). It follows that Erdogan’s campaign discourse is replete with religious sentiment and symbolism, variously expressing unity 10 and national (Islamic) identity 11 (under AKP rule). This lens is apparent in a campaign Instagram post (see Figure 3), 12 where Emine Erdogan, her husband, and party colleagues are seen breaking bread, accompanied by a warm lament for victims of the 7.8 magnitude earthquake in Turkey, which struck on February 6, 2023 (Figure 7):
Emine and Recep Erdogan Campaign Photo. Text: “On the blessed Night of Power (Laylat al-Qadr), we gathered with our citizens affected by the earthquake at the beloved table set in Üsküdar. We remembered our losses with longing and mercy, with the spirit of solidarity, giving each other strength and solace. On this blessed night, I pray to Almighty Allah for our prayers to be accepted, and for our tables of brotherhood to always remain.”
In contrast to Bolsonaro’s campaign discourse, however—which explicitly leans into the religious framing of political demands—recourse to religious symbolism and terminology is comparatively absent in Erdogan’s more policy-focused appeals:
Thank God, we have already brought the “real Spring” to Turkey in 2002, after years of crisis, instability and underdevelopment. We have strengthened this climate with the bold steps we have taken in the economy, democracy, fight against terrorism, rights and freedoms over the last 21 years. We raised the standards of our democracy by clearing the debris left by the tutelaries. We shelved the denial policies that did not value our people’s beliefs, language and culture, never to be brought to the agenda again.
13
The discursive structure of this Instagram post is practically the same as that of Bolsonaro’s emblematic YouTube campaign video of September 10th, 2022: an initial (though more perfunctory) appeal to God, followed by a neat series of political demands that conjures the specter of social collapse, all of which is accentuated by an explicit appeal to the people—the unity of whom shall be protected via a positive outcome at the ballot box. Much more so than “God” or “Islam,” Erdogan deploys “democracy” and “the economy” as the most important signifiers of competent authority rooted in stability and progress under his rule. Economic development, for example, strikes a decidedly nationalist tone:
Today, while we bring joy to 85 million people by commissioning our Black Sea natural gas, making our nation truly proud, of course we also upset some of them. We feel sorry for those who attribute their political success to the problems of our people . . . We upset the incompetent ambitious people who go around saying “there is no gas” . . . May our Black Sea natural gas be beneficial to our country and nation!
14
Likewise, Erdogan’s commitment to democracy projects a composite image of domestic unity and international prowess:
God willing, we will emerge victorious again from May 14, the most important elections in our history. So, who will win in these elections? . . . All our citizens of different faiths, including Christians, Jews, Assyrians, will win . . . Let us firmly protect our unity, solidarity, and thousand-year-old brotherhood in these lands . . . With God’s permission, the rest will come. May God preserve our brotherhood forever.
15
Unlike Bolsonaro, Erdogan is at pains to position himself as the incumbent rather than the outsider, which squares with his self-appointed role as protector of the realm. Nevertheless, his campaign is similarly structured along populist lines: Turkey’s social frontier is divided into two camps, accompanied by a chain of equivalence that ties the opposition—led by the Republican People’s Party (CHP)—to corruption, the (indirect) support of terrorism, the imposition of “LGBT values” and, ultimately, a splitting of “the people.” “Terrorism” acts as the most pivotal nodal point. It binds internal and external threats to Turkish unity on two bases. First, it is deployed to recall the (real) security threat posed by the Kurdistan Workers Party (PKK; who have perpetrated a substantial number of attacks on civilians within Turkey) and Erdogan’s capacity to keep the country safe:
Whoever tries to revive the old Turkey, which was under the threat of terrorism, will find us against them. Whoever puts guns in the hands of poor Kurdish children and tries to make the soldiers, police, and our people shoot bullets will find us against them. Whoever threatens the peace of our nation on behalf of the imperialist powers will find us against them. Whoever attempts to compromise your rights will find Tayyip Erdogan and the People’s Alliance against him . . . The darkness of terror will not fall on our cities at night.
16
Second, by virtue of welding the threat of “terrorism” to historical precedent and cultural memory, Erdogan can push spurious narratives that denigrate the CHP as a fundamentally corrupt, terrorism-supporting party, with links to shadowy, international agitators—all clear examples of campaign disinformation designed to discredit the opposition:
Before May 14, terrorist ringleaders shot a video from Qandil almost every day and publicly asked for votes for the CHP Chairman. They openly continued their support for this person after the election . . . Neither the CHP Chairman nor his friends at the table were ever offended or ashamed while organizing a joint rally with those who shouted “blood for blood, revenge for revenge” . . . They cannot hide the partnership they have established with separatists and FETO members in order to come to power and protect their seats. We will save our nation from the plague of terrorism that has been plaguing our country for the last 40 years, sucking the blood of our Kurdish brothers like a leech.
17
A synonym for Fethullah Gulen—the symbolic head of the Hizmet/Gulen international movement that strongly opposes Erdogan’s rule—“Pennsylvania” references the location of Gulen’s self-imposed exile in the United States to link the CHP’s internal operations with external interference. Officially designated a terrorist organization by Turkey (the Fethullah Terrorist Organization [FETO]) following a failed coup attempt on July 15, 2016, recourse to “Pennsylvania” permits Erdogan to raise the specter of internal division which would come by means of importation—that is, such a threat could never have been permitted to gain traction within Turkey under his rule. However, this threat can be accommodated by domestic and regional interlopers, including via Kemal/CHP and the PKK. This dynamic is well represented in a Facebook post on May 4, 2023:
The false spring that Mr. Kemal will bring means a dark winter for this nation. If Mr. Kemal brings it, it will only bring spring to Qandil and Pennsylvania. It brings spring to the terrorist nests in the North of Syria and Iraq . . . It will bring spring to the FETO traitors who dropped bombs on our nation on the night of July 15.
18
Finally, with the route between the CHP (internal)/terrorism(conduit)/Gulen (external) well established, Erdogan can thread through anti-LGBT disinformation as a symbol for the social (and secular) disorder that would be realized upon a CHP victory. By leaning into a global discourse that has gained much traction among authoritarian/populist leaders in recent years (Graff and Korolczuk 2022)—but adapting it as additional to the primary threat of security—Erdogan’s campaign discourse evidences a delicate balance between the internal and external; national and international; religious and secular. Universal narratives, such as “the imposition of anti-LGBT values,” must adhere to particular contexts and in Erdogan’s campaign discourse, “religious disinformation”—classically understood—is conspicuous by its absence. Instead, it is the familiar threat of “terrorism” that looms largest:
Those gathered around the table in front of us [members of the CHP] were not chosen randomly. Each of these people are assigned different missions in the project of blocking Turkey . . . One of them is there to overshadow the ambitions of the separatist organization, which is supported and strengthened in Northern Iraq and Syria, on our country. He is supposedly a nationalist, but he does not speak out against the statements made by Qandil and the bold campaigns carried out by the PKK. Another one is on the table to overshadow the LGBT imposition, which is the biggest trend threatening the future of Western countries. If you ask, they would say they are nationalists, but they do not speak out against projects that undermine the national values of our nation.
19
Discussion and Conclusion
Our methodological approach was crafted to discern the religious discourse of Erdogan and Bolsonaro as part of their curated election campaigns. We formulated a religious lexicon based on the candidates’ own speeches and scrutinized all disinformation linked to this lexicon, whether it constituted false religious information or not. Contrary to our expectations, we did not encounter “religious disinformation” per se: the data did not align with the concept as utilized by authors such as Ahmed (2018), Alimardani and Elswah (2020), Al-Zaman (2021), and Al-Zaman and Noman (2023). In our analysis, we did not identify misleading religious guidance, misinterpreted scriptures and/or religious records (such as the Bible), or misrepresentations of divine knowledge or authority (Alimardani and Elswah 2020).
The statistical analysis of social media posts from the 2022 Brazilian and 2023 Turkish presidential campaigns revealed distinct patterns in the use of religious rhetoric. For Bolsonaro, religious content was significantly more prevalent on platforms such as YouTube, and YouTube TV, while significantly less prevalent on Twitter and Telegram. This suggests a strategic deployment of religious messaging on these platforms to consolidate authority. Engagement metrics further underscore the differences between platforms. Bolsonaro’s religious posts garnered significantly higher engagement on YouTube TV, highlighting the platform’s effectiveness in amplifying his religious rhetoric. Conversely, Erdogan’s religious content showed a positive engagement effect overall, yet interactions on individual platforms like YouTube, TikTok, Twitter, Instagram, and Telegram were notably lower for religious content compared to nonreligious posts. This dichotomy suggests that while religious messaging was a central theme in Erdogan’s campaign, it did not uniformly translate to higher engagement across all platforms.
These findings illuminate how both leaders instrumentalized religion in their digital campaigns, albeit with differing strategies and outcomes. Bolsonaro’s broader use of religious rhetoric across multiple platforms aligns with his identity as a messianic leader, leveraging a narrative of divine support to strengthen his political authority. Erdogan’s more selective use of religious content, combined with his focus on democracy and economic development, reflects his established religious credentials and long-standing political strategy.
As our qualitative findings rest on textual filtering, the relationship between these leaders and religiosity manifested through gestures and the use of iconographic elements was addressed tangentially. Thus, future research based on other methodological approaches is still needed to thoroughly address the imagery dimension of digital populist communication (Mendonça and Caetano 2021). Given the focus on presidential official campaign discourse disseminated across platforms, other aspects of the electoral race, such as conversations, social mobilization, influencer activity and in-person campaign, on rallies or religious events, were not included. Ethnographic approaches as well as in-depth surveys would certainly contribute to broadening the scope of conclusions about political and religious communication practices that take place in the digital sphere during the campaign.
By allowing the data to lead and tracing, thereafter, how religion and disinformation coalesced under a rubric of populist articulation, what surfaced, instead, was how religious rhetoric and disinformation tended to orbit one another, as opposed to being conjoined as a novel category. Indeed, our findings around the instrumentalization of religious rhetoric toward clear political ends chimes with what Domke and Coe (2010) have labeled the “God strategy”: a phenomenon observable in US politics—and political campaign discourse—since the early 1980s, at least. Consequently, the concept of “religious disinformation” seems, at present, insufficient to capture the role of religion in disinformation strategies and/or political campaigns. Our findings dictate that a discourse may be replete with religious rhetoric and disinformation without constituting “religious disinformation.” To that end, we argue for more heterogeneous interrogations of how religion and disinformation interact in diverse political settings. Perhaps these findings—gleaned from a comparative analysis of countries belonging to the “Global South”—can be re-tested in North(west)ern political theaters, where the assumption of secularity is much more ingrained.
Nevertheless, our findings must be maintained in their proper context. They, too, are insufficient to move the dial on religious disinformation in the “Global South” in any absolute sense: our analysis is exploratory, and our case study foci are relatively narrow. In order to further interrogate the interplay between religion and disinformation in political campaigns, questions remain on matters including the spread of religious disinformation across Encrypted Messaging Apps (EMAs)—such as WhatsApp and Telegram—where unofficial content can spread at speed and where dynamics of participatory disinformation become far more apparent (Santini et al. 2021; Veilleux-Lepage et al. 2022). Furthermore, as data was collected postelections, the candidates and parties might have deleted posts, and the amount of unavailable content is unknown. For instance, we know that collective belief in demonstrably false information is driven by an array of factors, including the need for community belonging, epistemological certainty and ontological security (Fitzgerald 2022). However, to the extent that official campaign discourse is meticulously crafted, it necessarily takes into account the needs—and expectations—of the audience (Atkins 2022; Jensen 2017). To that end, if there is little or no recourse to religious falsehoods in the official discourse of Erdogan and Bolsonaro, perhaps it is because the religious communities to whom they speak shape the limits of faith-based falsehoods: an organic form of faith-checking, which requires further interrogation.
Supplemental Material
sj-docx-1-hij-10.1177_19401612241285660 – Supplemental material for Heaven Holds a Place for Those Who Pray: Instrumentalizing Religion and Disinformation the 2022 Brazilian and 2023 Turkish Presidential Campaigns
Supplemental material, sj-docx-1-hij-10.1177_19401612241285660 for Heaven Holds a Place for Those Who Pray: Instrumentalizing Religion and Disinformation the 2022 Brazilian and 2023 Turkish Presidential Campaigns by James Fitzgerald, Ayse D. Lokmanoglu, Debora Salles and Rose Marie Santini in The International Journal of Press/Politics
Footnotes
Acknowledgements
We would like to thank Bruno Mattos for his invaluable support, as well as the reviewers and editors for their meticulous review and constructive feedback.
Data Statement
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
Supplemental Material
Supplemental material for this article is available online.
Notes
Author Biographies
). She leads as Principal Investigator the research projects “Opinion Machines: Computational Propaganda, Contagion and Misinformation in Online Social Networks” (2020–2024) and “The Media Ecosystem Dynamic Regarding Environmental Issues, Land Use, Indigenous People’s Rights and Support for Climate policies in Brazil” (2020–2025). She is Open Society Foundation, CLUA, Ford Foundation and CNPq (Brazilian National Council for Scientific and Technological Development) Fellow.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
