Representing “The People”: What Can Social Media Images Reveal About Populist Propaganda in Brazil?

Abstract

This study explores the visual content posted by far-right populist parliamentarians in Brazil, asking how “the people” are represented through imagery. Scholars have previously argued that populists have failed to foster the voice of “the people” in Latin America, but have they also failed to signify and cultivate an authentic visual representation of them? To answer this question, we collected the social media imagery posted by federal deputies of the 56th Legislature, a period characterized as one of illiberal backlash in Brazil. We employed visual analytics as a method of data analysis, demonstrating that clustering images according to their color similarities facilitates the study of key themes of populist communication, namely: the leader’s bodily performance, nationalism, the “enemy,” and “the people.” Besides this methodological contribution, our findings suggest that parliamentarians played a crucial role in enabling and sustaining far-right populist discourses. These politicians defined “the people” solely through the battles set in motion at the level of political affiliations. The voices of those claimed to be represented by them were diluted in visuals of mass demonstrations, which provided the primary occasion for their incorporation into populist communication.

Introduction

The surge of far-right populism in Brazil has raised concerns about how parliamentarians interact with their constituents online. These mediated interactions have been at the center of recent political turmoil: the Supreme Court ordered the arrest of a federal deputy for instigating violence against its justices on YouTube (Rocha et al., 2021), the Electoral Supreme Court ordered the suspension of parliamentarians’ social media accounts for disseminating false claims of electoral fraud (Ribeiro, 2022), and the Senate Inquiry Commission (No. 1.371, 2021) investigating the mishandling of the COVID-19 pandemic recommended that charges should be brought against members of the Federal Chamber for their disinformation campaigns. These cases suggest the need to seriously consider how parliamentarians speak with their constituents. As studies on populist political communication have focused primarily on Brazil’s most prominent politician, former president Jair Bolsonaro, less is known about how his political allies can enable and sustain his populist agenda.

This research examines how far-right populist parliamentarians used social media in times of illiberal backlash in Brazil (Power & Hunter, 2019). Scholars have previously argued that populists have failed to foster the voice of “the people” in Latin America (Waisbord & Amado, 2017), but have they also failed to signify and cultivate an authentic visual representation of them? To answer this question, we investigated the social media imagery posted by parliamentarians in the course of their mandates. Specifically, we identified the most far-right populist federal deputies of this legislature and collected all the images they posted on Twitter, a procedure that yielded a sample of 18 politicians and 16,740 images. We then employed visual analytics (Rogers, 2021) as a method of analysis, demonstrating that clustering images according to their color similarities facilitates the study of populist visual frames. Our findings suggest that parliamentarians play a crucial role in populist projects. In their attempts to (re)construct an ideological domain, these politicians defined “the people” solely through the battles set in motion at the level of political affiliations. The voices of those claimed to be represented by them were diluted in visuals of mass demonstrations, which provided the primary occasion for their incorporation into populist communication.

Literature Review

Populist Political Communication

The centrality of populist communication has long been recognized in Latin American politics. Pedro Ernesto, described as the head of Brazil’s first populist movement in the 1930s, built his popularity from innovative communicative techniques that mobilized political participation, particularly his use of radio in the 1930s (Conniff, 1981). His techniques would later be replicated by Getúlio Vargas, who reached out to the poor in the first city-based mobilizations of mass-electoral support (Carvalho, 1998). Indeed, scholars tend to agree that the first manifestations of this political logic coincided with the rise of mass media publicity (Canclini, 1997). But as populism historically relied on a media apparatus inaccessible to many, contemporary populist actors rely on a new media technology: social media platforms. This section explores how the literature contributes to our understanding of populist communication, and the role of parliamentarians in it.

Populism is a notoriously difficult concept to define. This difficulty is derived from the concept’s various facets: from Ernesto—who Vargas imprisoned after veering toward socialism—to Jair Bolsonaro, populism has been used to describe significantly different political projects. Yet, Laclau (2005a) saw the flexibility of the concept as its defining characteristic. He argued that populists cannot be branded simply because of the demands they claim to represent, but because of how they articulate these demands discursively. This articulation relies upon logics of equivalence and difference to constitute a unified collective identity, “the people,” as an antagonistic force against their opponents.

A key aspect of populism is its discursive construction of “the people.” In populism, as Panizza (2005) explains, the relationship between the form (“the people” as a signifier) and content (what is signified) is understood through an antagonistic mode of identification. Populist leaders articulate a plurality of unfulfilled demands into seemingly coherent totalities, transforming grievances into “elements in a chain of equivalences in which they only have in common the relation of antagonism itself” (Panizza, 2005, p. 6). Jair Bolsonaro’s ascendancy to power illustrates this. Bolsonaro established equivalential relations between demands for public security, gun access, and intensified state punitiveness—demands which, despite having no necessary connection between them, were aggregated into equivalential chains (Silva & Rodrigues, 2021). Building upon this, he blamed “leftists” ruling parties for ignoring the needs of the people, transforming the nature of each particular demand: they cease to be isolated requests and acquire new meaning within a wider populist discourse.

The story of the 2018 election is complicated, but what is important is that Bolsonaro, and his allies, sought to inscribe a range of demands into a new, unapologetic (Pierucci, 1987) right-wing populist identity. The cohesion of this populist project depended on the appropriation of signifiers (“the people,” “the nation”) which stood in for the equivalential relation as a whole. The resulting political identity operated as an empty signifier, ¹ an object of investment which transcends its particularistic contents: in Bolsonaro’s case, calls for loosening gun regulation started to form a relation of contiguity with “being a patriot,” a “good Christian,” and so on. One political stance became a metaphor for another, crystalizing an apparently unified collective will.

Populism draws part of its force from recent transformations in our media landscape. Today, populists are able to interact with their supporters on a continuous basis using social media platforms, which allow them to bypass traditional media gatekeepers (Engesser et al., 2017). The network logic of these platforms enables virtual connections with like-minded individuals, facilitating the dissemination of their messages (Stier et al., 2017), which may appear more credible by virtue of likes, shares, and other instruments of false spontaneity (Woolley & Howard, 2018). Moreover, social media algorithms create structural incentives for antagonistic content that can foster this political logic (Rathjea et al., 2021). Reflecting upon these technological changes, Foster (2022) suggested that we should examine contemporary political communication as part of a culture of iteration, that is, a context where information’s persuasive power depends on the repetitions and reinforcements afforded by digital media.

If iteration has become a crucial aspect of communicative processes, the continual presence of political actors on their audiences’ feeds can signal a new form of the contest over the interpretation of our state of affairs. With the weakening of traditional bases of validating information (Bennett & Pfetsch, 2018), politicians have increasingly taken up that role. And in helping their audiences make sense of a complex state of affairs, they repeatedly reinforce their political identities (“this is what the left is hiding from you”; “another case of attack against our Christian values”), effectively binding events to ideological discourses. However questionable these discourses may be, they help to “classify out the world in terms of immediate political and moral values” (Hall, 2010, p. 130).

The decline in professional journalism contributed to the consolidation of these new communicative paths. Latin America’s media systems have been historically criticized for their concentration of ownership and close ties with state officials (Waisbord, 2000). In Brazil the dominance of a few news organizations became increasingly contested after major protests in 2013 (Arnaudo, 2018) and by retrenchment in their advertising revenues (Fabrini & Wiziack, 2020). In this context, social media platforms, where information is less verified (if at all) have become the main media source of news for Brazilians. This is especially concerning as, on these platforms, political leaders are often the main spreaders of disinformation (Reuters Institute, 2020). Relatedly, members of the press perceive politicians as competing with them over audiences’ trust, as populists openly claim that mainstream journalism has become dispensable (Mont’Alverne et al., 2021).

Besides transformations in our media landscape, populism also draws its force from citizens’ discontent with democratic institutions. In a time when political parties failed to renew their connections with voters (Hunter & Power, 2019), Bolsonaro and his allies sought to present themselves as an alternative political force. This alternative, they claimed, was the only one that truly represented the Brazilian people. Populists often seek to establish a sense of “virtual immediacy” (Arditi, 2005) with their constituents, ascribing the interests of their leaders to the citizenry as a whole. By taking for granted these representative bonds, populists subscribe to what Müller (2014) called a moralistic imagination of politics. When this is taken to its extreme forms, this mode of representation does not require the “input and continuous influence by citizens divided amongst themselves” (Müller, 2014, p. 487). Consequently, the moralistic imagination underlying populism oscillates between being a restorative force (Silva & Vieira, 2019) and an affront to liberal democracies (de la Torre, 2017).

These criticisms raise a productive line of inquiry: are populists in Brazil actually promoting representative mechanisms that summon the presence of “the people”? If they indeed subscribe to a moralist imagination of politics, we can expect a negative answer to this question. We posit that populist communication fails to promote and cultivate authentic representations of “the people,” reproducing instead a subaltern form of citizenship. Drawing on Santos’ (2018) work, we use the concept of subalternity to describe how certain constituents, despite being referenced and argued about in representative institutions, are accorded no discursive position from which they can speak for themselves. In other words, citizens are seen as objects of, but never the subjects of their own political interventions. Populist communication can reproduce a “culture of silence” (Freire, 1970) by constituting the unity of the people “from the balcony of the government palace” (Canclini, 1997, p. 480), or in contemporary terms, from behind the screens of their media devices. Despite unparalleled access to online spaces, citizens would still lack the discursive power (that politicians have) to speak on their behalf, suggesting a pervasive subalternity even on social media platforms. This is the first line of inquiry advanced by this study.

• Hypothesis 1: Populist political communication will fail to promote and cultivate an authentic representation of the people, reproducing instead a subaltern form of citizenship.

In sum, benefiting from media and democratic crises, populism has risen as a major topic of concern in recent years (Engesser et al., 2017; Hunger & Paxton, 2022). Yet, we need to move beyond the assumption that populist communication can be studied solely through the practices of its leader. Laclau’s conceptualization of populism has to be expanded to include “the presence of strategically placed agents who can construct and deploy empty signifiers to advance their projects” (Howarth, 2004, p. 262). Populist projects benefit from legislative coalitions held together by shared ideologies (Zucco, 2011) and congressional rules (Power, 2011). We know that the organization of parliamentarians’ cabinets has adapted significantly with the rise of social media (Carlomagno et al., 2019), and these politicians have profusely used public resources in mounting propaganda campaigns (Mello, 2020). The scholarship clearly reveals the significance of digital media in the legislative arena, but it does not directly address how these platforms are related to the resurgence of populism in Brazil. For this reason, our understanding of how populist discourses are enacted, or even strengthened, by parliamentarians remains patchy.

Social Media Framing of Populism

As a “bridging model for media research” (Reese, 2007), framing theory complements Laclau’s interpretation of populism. If populist discourses are defined by the constitution of a popular identity, “the people,” in confrontation with an antagonistic other, these identities can take different contents and shapes, so we expect that the media also play a constitutive role in fixing its meaning. Framing theorists have been traditionally concerned with how texts help us make sense of a social reality by not only describing, but by actively offering an interpretation of it (Entman, 1993). Yet, the theory has been approached by distinctive epistemological traditions (Van Gorp, 2007) and applied to examine different communicative processes (cf. Meraz & Papacharissi, 2013; Porto, 2004). Thus, it is important to state clearly what definition scholars have in mind when working with this theory, not just to situate themselves in relation to the existing literature, but also because different definitions may lead scholars in different directions (Bock, 2020).

As we are interested in how parliamentarians signify the identity of the people, we apply Reese’s definition of frames as “organizing principles that are socially shared and persistent over time, that work symbolically to meaningfully structure the social world” (2001, p. 11). Framing serves populist political projects by emphasizing what popular demands deserve our attention and what constituents are impacted by them, thus being marked by constitutive inclusions and exclusions. This operation is consistent with Laclau’s understanding of the boundary work necessary to constitute “the people.” Notably, Laclau argued that requests, to be politically meaningful, must become part of wider discursive formations. Requests are not translated politically on their own, and may acquire a populist configuration depending on the frames used by political leaders in specific historical conditions.

So, what is distinctive about populist framings? We previously noted that populist politics is marked by a dissimulation of the distance between representees and representatives. To achieve that, populists’ self-presentation as standing in for “the people” often relies on personalistic leadership (Panizza, 2005) and a distinctive bodily grammar (Casullo, 2018), constructed through the display of the leader’s gestures, clothes, and other markers that portray the leader as a common person. We also noted that populism depends on antagonistic modes of identification, so messages that criticize, insult, and ostracize opponents are prevalent in their communicative style (Jagers & Walgrave, 2007). Indeed, scholars have suggested a number of features that distinguish populism not only as a political logic but also as a style of communication. These features include tabloid-like language (Canovan, 1999), the negativity of their messages (framing reality as a continuous crisis, evoking outrage and resentment, Engesser et al., 2017), and fixation with news coverage (Waisbord & Amado, 2017).

Behind this communicative style is the intent of speaking directly with “the people,” bypassing the press and other gatekeeping institutions. Populists are known for their presumption of enjoying a privileged rapport with their followers (Engesser et al., 2017). Perhaps unsurprisingly then, populist politicians often seek out nontraditional approaches to communicating with and about the public. One exemplar of this is Bolsonaro’s encounters with hardcore supporters on the doorsteps of the presidential residency which have been recorded and widely shared online; and in many instances, these encounters are meant to substitute open press conferences (Phillips, 2020). Social media platforms, therefore, allow populist leaders to circumvent the reliance on traditional news attention–or, better yet, they help populists garner more news attention.

Though politicians rely on a collection of platforms to garner public attention, among these, Twitter emerged as a uniquely important one. One reason for this is the composition of its user base, which includes prominent figures such as other politicians, journalists, activists, and political groups (Belotti et al., 2021). While less than 10% of Brazilians are on this platform (Kemp, 2022), Brazilian journalists rely heavily on Twitter for their profession (Peres-Neto, 2022). Twitter’s function as the preeminent platform for journalists has been documented both in Brazil (Fraga, 2012) and in other democracies around the world (Rauchfleisch et al., 2017; Verweij & Van Noort, 2014). Thus, if politicians want to garner news attention, Twitter may enable them to do so (Wells et al., 2020). Furthermore, content that politicians (or their staff) post on Twitter tends to be cross-posted to other platforms, creating a coordinated and complementary strategy (de Barros et al., 2021). Indeed, much of the content we document in this article is also circulated on other digital spaces. ²

Given politicians’ use of Twitter to garner media and public attention, it is worth considering how populist communication is reproduced through their tweets. In Latin America, scholars have sought to investigate whether this platform confers a unique, nontraditional flair to this form of communication. But instead of favoring horizontal interactions between populists and “the people,” Waisbord and Amado (2017) found that populists used Twitter to bolster the information power of the presidency. As a result, the authors contend, populists’ use of Twitter fell short of its interactive potential, an argument that corroborates with our subalternity hypothesis. However, this scholarship has focused primarily on government leaders (Manfredi-Sánchez et al., 2021), from Trump (Wells et al., 2020) to Bolsonaro (Fontes & Marques, 2022), leaving aside how other political figures in other levels of state participate in populist communication.

Extending this theoretical framework to the study of parliamentarians’ communication requires acknowledging that these actors as actively involved in framing public debates (Recuero et al., 2019). Following Entman and Usher’s (2018) calls for more studies on the new paths of communication between elites and the public, we explored how these politicians influence the flow of political information, sometimes akin to whom they described as rogue actors: that is, those whose actions are seemingly unconstrained by the norms which used to dominate political communication. Other scholars expanded the theory beyond news coverage, suggesting that framing processes are today performed by a range of social actors (Meraz & Papacharissi, 2013). For example, Dias et al. (2021) described populist framing mechanisms of grassroots right-wing organizations, suggesting that antagonistic messages enabled them to form a united front against a common enemy, even if their motivations and prognostics differed. Yet, considering the widespread discontent with representative politics in Brazil, we decided to focus on actors who contributed to delegitimizing its democratic institutions (Phillips, 2021).

Visual Frames of Populism

Parliamentarians increasingly participate in framing processes through social media imagery. Digital media and their associated technologies have afforded these actors the possibility of easily circulating visual narratives, contributing to the formation of popular identities. Visual frames enable certain objects to be more easily invested with meaning, thus extending Laclau’s (2014) postulate that “representation will always be figural” (p. 123). We previously noted how rhetoricity is at the core of the constitution of “the people”: through a metaphorical operation, this popular identity is constructed by establishing equivalential relations between disparate grievances and demands. Visual frames contribute to populist discursive articulations by means of transforming certain objects (the figure of the leader, for example) into the name of an absent universality (the body of the leader is invested with radical anti–status quo feelings). These discursive articulations occur in a way that is more easily interpreted and recalled by audiences, as visual stimuli tend to elicit stronger emotions from them.

In populist communication, the figure of the leader becomes an object of symbolic investment, transforming their bodily self-presentation into a source of representation. The bodies of populist leaders work symbolically to personify their bond with “the people” (Casullo, 2018). By presenting themselves as ordinary people, populists’ bodily images carry the presence of their followers in a transgressive manner. Mendonça and Caetano (2021) find that Bolsonaro employs a hyperbole of this transgressive style. His visual performance confers a down-to-earth quality to the liturgies of his mandate, “as if his body misfits the extraordinariness of the presidency itself” (Mendonça & Caetano, 2121, p. 218). Bolsonaro’s social media imagery portrays a leader whose body displays ordinariness while the leader simultaneously breaches official protocols, as exemplified by numerous photographs where he wears football jerseys in official events. Similarly, Donald Trump’s self-presentation has also been described as a typical case of populist visual communication by virtue of his norm-breaching performances, defiance gestures, and attention-getting expressions (Bucy, 2022). Bucy et al. (2020) give one step further, claiming that such demonstrations of a transgressive style more easily resonate with voters and boost metrics of social media engagement.

These examples highlight how the articulation of populist discourses relies on the bodily display of populist leaders. For Bucy et al. (2020), it is their transgressive performances that distinguish populist visual (but also verbal and tonal) self-presentation. For Casullo (2018), these leaders combine ordinariness, exceptionality, and power in diverse ways, but these three dimensions are constitutive of populism’s aesthetics. It remains an open question, however, if political allies contribute with these visual performances by merely reiterating the leader’s images or whether they reproduce the same form of symbolic investment through their own self-presentation. Moreover, we still have much to learn about alternative objects of investment that may operate in tandem with bodily displays in populist communication, including “the people,” “the nation,” and other empty signifiers. But what is clear in this body of literature is that visual communication plays a key role in the enactment of populism (Bucy, 2022), having its own explanatory value.

The importance of visuals has been highlighted by studies on media and cognition. Images tend to generate stronger effects than texts on opinions and behaviors (Powell et al., 2015), which suggests that visual frames more easily enact cognitive schemata. Relatedly, Domke et al. (2002) suggested that images offer interpretative frames that may be more easily processed and recalled by audiences (see also Huang & Fahmy, 2013). Besides more easily capturing attention and facilitating the interpretation of messages, visual stimuli also elicit stronger emotional responses from audiences (Brantner et al., 2011).

Our study seeks to contribute to this literature by exploring the role of sensory data, particularly color patterns, in populist visual frames. Scholars have previously examined the use of sensory data in conspiracy messages (Chen et al., 2022) and misleading information (Cao et al., 2020) using computational methods. They found that colors tend to be manipulated to persuade audiences and elicit emotional responses from them, such as when darker colors are used in conspiracy videos to provoke fear and anxiety (Chen et al., 2022). If this insight can be extended to the study of populism, we expect to find associations between color palettes and key objects of populist articulation, thus painting “the people” in ways that connote patriotism for example. To verify that, our second line of inquiry posits that colors can help to encode key objects of populist communication, serving as heuristic devices scholars can use to learn about the populist visual frames.

• Hypothesis 2: Colors enable populists to more easily encode key topics of their political communication, serving as a heuristic device in the study of populist visual framing.

The reason why colors have a unique role in visual frames is that, for articulations of abstract ideas, colors can be used as a way of priming a range of emotional and interpretative responses (Elliot & Maier, 2014). Examples of that can be found in Guilbeault et al.’s (2020) study, which demonstrated that emotions and abstract concepts tend to be represented through analogous color associations (e.g., images referencing “hate” were connected with reddish clusters on Google Image search results). They concluded that colors encode affective and logical associations, an insight that is of key importance considering this paper’s method (which relies on creating image collections assorted according to their colors, as the next section explains) and object of study (articulations of other abstract concepts, such as “the people” or “the enemy” in populist discourses). This helps to explain why far-right populists tend to paint their opponents in red while presenting themselves with patriotic colors—often “inflicting” (Hall, 2010) this content with religious connotation, as exemplified by Figure 1.

Figure 1.

“Choose Your Side”: Tweet with the caption “the struggle between good and evil has never been so evident! #BolsonaroPresidentUntil2026.” The image presents the former President Lula da Silva holding the hand of the devil. On the other side, Bolsonaro is walking alongside Jesus Christ with followers carrying national flags and wearing green and yellow clothes (Zambelli, 2021).

Methods

This research examines the visual communication of the most far-right populist parliamentarians on Twitter. These parliamentarians have been at the center of the political controversies mentioned in this paper’s introduction, allegedly being associated with new forms of computational propaganda (Mello, 2020; Woolley & Howard, 2018). ³ With the support of Capture and Analysis Toolkit (4CAT), we collected all tweets these politicians posted during the 56th legislature, a time period which comprises the years of 2019, 2020, and 2021, excluding the electoral cycle of 2022. We employed visual analytics to inductively examine the most prevalent visual frames in our dataset. This method consists of creating image collections and arranging them according to their aesthetic properties. Based on this procedure, researchers can more easily identify patterns and changes over time in large collections of images (Rogers, 2021), thereby learning about the phenomenology of platform vernaculars and their storytelling capacities (Pearce et al., 2020).

Our sample procedure consisted of three main steps. First, we identified the members of the Federal Chamber of Deputies most aligned with the surge of the far-right in Brazil. To do so, we used the percentage of votes in favor of the government’s bill propositions. ⁴ With 21 parties represented in the lower house, frequent changes in their denominations, and shifting coalitions throughout the legislature, voting patterns on the congressional floor can offer a relatively simple and accurate identification of where parliamentarians fall across the ideological spectrum. The threshold to be included in this preliminary list was 90%. A total of 249 parliamentarians met this condition. ⁵ Second, politicians who did not have a strong presence on Twitter were excluded. In alignment with the Senate Inquiry Commission (No. 1.371, 2021), we operationalized “strong” presence as having 50,000 Twitter followers or more (as of November 2, 2021), yielding a reduced sample of 24 parliamentarians.

Finally, we selected the most populist parliamentarians for our final sample. This selection was based on a textual analysis whereby we counted the number of occurrences of potentially populist terms in each dataset; to conduct this analysis, we use the R package “tidytext” (Silge & Robinson, 2016). This terms list is adapted from Ricci et al. (2021). The list consists of terms that can both signify “the people” and “the enemy” (see Appendix B). We followed a dictionary-based approach in our textual analysis, thus breaking down our dataset into word tokens (with each token being given a “people” score or “enemy” score of either a 1 or 0, depending on the inclusion of the concerned category or not). The scores were aggregated at the post level and subsequently at the user level. Next, we compared the scores across the dataset. This comparison took into consideration the variability of the number of posts for each profile, as we averaged the number of occurrences per tweet. The averaged “people” and “enemy” scores were combined to yield the final sample, consisting of 18 parliamentarians who scored at least 0.91 per post (maximum score Eduardo Girão 1.92 per post). There was a considerable gap between the active adopters of the populist frame and those who did not, with six parliamentarians scoring less than 0.68 per post. In total, these parliamentarians posted 16,740 images in the course of this study’s time frame.

It is worth noting that these methodological strategies cannot fully capture the meaning of any of the political constructs being referred to in this paper. In the same way that a far-right political orientation cannot be defined solely based on congressional votes, we are not claiming that populism can be understood as a compilation of linguistic terms. But, as suggested by Laclau (2005b), the challenge is not to identify whether politicians are populists or not, but to what extent populist logics dominate their discourses. In this sense, the value of our approach lies in identifying one way in which both political phenomena become manifest, thus grounding our sampling procedure in empirical data.

The images embedded in our final datasets were analyzed using ImageSorter (Visual Computing, 2018). This software operates as a browsing application which produces a two-dimensional grid with all images collected. It groups these images together according to their color similarities (alternatively, it can also assort them based on size or date taken). Once the images were plotted and assorted according to similar color palettes, as Figure 2 shows, we examined the different clusters in search of prevalent themes and changes over time. Indeed, we believe that the “kaleidoscopic” appearance of the image mosaics helps us better understand the appeal of populist political communication. This methodological procedure allows us to think of political communication beyond discrete propaganda campaigns or populist themes (such as images with bodily display, as in Casullo, 2018 study), expanding our analysis to the modes of expression of a particular coalition of political actors. By following this approach, our goal was to identify, inductively, the most common visual frames in our samples’ political communication.

Figure 2.

Image plot of Twitter sorted by color in ImageSorter. From left to right, these mosaics contain all images posted by parliamentarians in our sample in 2019 (totaling 4,469 images), 2020 (6,732 images), and 2021 (5,539 images).

Results

This section describes the visual messages posted by far-right populist parliamentarians in the course of their legislative mandates. This study is based on three image collections produced by ImageSorter (see Figure 2), covering the 3 years of this paper’s period of study. Based on these mosaics, we identified five homogeneous clusters. The first two included images associated with what we describe as knowledge production, as they presented interpretations of Brazil’s state of affairs. The second cluster concentrates, on the one hand, on photographs of official occasions and liturgies of the mandate, and on the other, selfies that reveal these politicians’ personal lives. In the red cluster, we found images focused on antagonism with political opponents. Finally, the green cluster outlines how nationalism has ceased to be a signifier solely evoked to defend the government’s agenda, more recently being enacted to further entrench divisions between elites and the “people.”

The analysis proceeds in the following way: first, we describe each cluster, highlighting and comparing patterns in the data for 2019. Next, we investigate how the clusters have undergone transformations, ⁶ in content and size, laying the ground for the subsequent discussion about how populist communication has evolved throughout this legislature.

Black and White Clusters: Knowledge Production

As the dominant cluster in all image collections, the white cluster consisted of screenshots from mainstream media headlines, other politicians’ Twitter accounts, and official documents. These first two categories of images predominantly referenced government actions, with occasional attacks against oppositional figures for blocking the advancement of their agenda. Similar to Recuero et al.’s (2019) study, this finding signals a persistent attempt by politicians to curate media content, respond to it, and share their own views about ongoing public debates. And although this finding aligns with Waisbord’s (2018) argument that populist communication tends to be fixated on news coverage, this did not lead to an embracement of the wisdom of “the people,” who remained nevertheless absent as sources of content from this cluster—and the next. Therefore, the dominance of this cluster suggests that far-right populist parliamentarians are strongly invested in providing interpretations of ongoing controversies.

Both black and white clusters (Figure 3) displayed images conveying a sense of unmediated access to the reality of power. In the latter, we identified numerous images containing authoritative documents, with the official header of the Chamber of Deputies, stamped and signed by parliamentarians. In the former, parliamentarians reproduced the Chamber’s scoreboard for voting sessions, claiming victory over the approval of legislative bills. Both types contain a distinctive appeal to truth, but the scoreboard pictures also fit an antagonistic framing of politics.

Figure 3.

White and black clusters of the 2019 image mosaic, created with ImageSorter (Visual Computing, 2018).

Alongside images from inside the Chamber, the black cluster displayed several quotes of well-known figures, from Winston Churchill to Olavo de Carvalho (considered Brazil’s intellectual mentor of the far-right). Another recurrent source was the Bible because of Marco Feliciano’s tweets, a pastor who belongs to the evangelical caucus in parliament. Yet the importance of these textual inscriptions is not solely derived from their references. These quotes raise a range of themes, from Christianity to the value of truth, which assert moral codes and ingrained attitudes. Through these authoritative excerpts, this cluster substantiates a cultural practice in which what is at stake is not just the provision of “information but confirmation, not to alter attitudes or change minds but to represent an underlying order of things” (Carey, 2008, p. 15). In other words, it indicates an expanded scope of parliamentarian communicative activity which is, beyond the transmission of (dis)information, engaged in the ideological organization of everyday life.

In the two subsequent years, the white cluster increased in size. The content maintained the same frame of reference: screenshots of news media, tweets from other politicians, and official documents. It is worth noting, however, that other oppositional politicians started to appear in this cluster, with more of their tweets being replicated through these images. The black cluster had no perceptive changes in its size. Interestingly, we observed the inclusion of quotes from the president and his secretaries in this cluster. Both trends suggest that Brazil’s political class has become more auto-referential in the course of this legislature.

Beige Cluster: (Extra) Ordinary Authority

The beige cluster occupied the second-largest area of the 2019 image collection. In this cluster, we find numerous photographs taken alongside other politicians, particularly president Bolsonaro and ministers of state. As seen in Figure 4, these photographs have their basic value derived from the register of what alliances are being formed and maintained. The connection between Legislative and Executive power is made visible for the audience to witness. In this way, this cluster gives further evidence on the use of Twitter as a platform of interaction between political elites (de Barros et al., 2021).

Figure 4.

Examples of images in the beige cluster of 2019.

But these photographs call for further elaboration. Besides staging political alliances, this cluster deals with bodily displays of power (Casullo, 2018). The repetition of poses, luxurious clothes, and ostentatious body language assert the authority of these parliamentarians. In their invocation of power, these photographs also appear to reveal the interior of the state institutions: with a national flag in the background, they are set in the remote offices and halls of Brasília. This form of representation “before the people” (Habermas, 1991) creates a distance between these politicians and their audiences: it signifies their authority and establishes an extraordinary status attribute.

And yet, though this cluster plays with markers of distinctiveness, it also consists of numerous selfies reducing the distance created by other photographs. As the right column of Figure 4 illustrates, these selfies bring back a sense that these politicians are just like everyday people, conferring a sense of ordinariness. Therefore, although populist identification relies on the constant drawing of boundaries between the elite and “the people,” here we observe how parliamentarians capitalize on the ambiguity of these categories: they simultaneously mask and reclaim the reality of power.

The beige cluster had a considerable decrease in size after 2019. Ceremonial images have become less frequent, but many of them are still present throughout this legislature. Similar to other areas of the image collections, the core of the beige cluster has started to display photographs of mass demonstrations, an important development detailed in the next sections.

Red Cluster: Antagonizing with the Enemy

Against our expectation that far-right populist parliamentarians would post a large number of reddish images attacking their opponents, 2019 had a modest red cluster. It also included references to a range of events that occurred that year, reducing its initial homogeneity. However, the cluster was not lacking antagonistic content altogether (see Figure 6, Appendix). It presented several images ridiculing former president Lula, imprisoned for almost the entire duration of 2019. It also criticized and ostracized unions and social movements. In common, these antagonistic posts shared a clear underlying message: under this administration, the people can celebrate that the left was ousted from power.

That message dramatically changed in 2020 and 2021. Celebrations gave way to images targeting governors and other political figures in condemnation of their “hypocrisy” throughout the COVID-19 pandemic (for example, with photographs of them without wearing a mask, against their own health recommendations). The 2021 red cluster further radicalized by means of attacks directed toward media organizations and protests against the president. Regarding the latter, several images posed a contrast between demonstrators in favor versus against the government, articulating whiteness and femininity to their political project (see Figure 7, Appendix A). Although more studies could help substantiate this claim, we noted how visual representations of women pro-Bolsonaro tended to be white, posing in an orderly manner, wearing green-and-yellow clothes; those against him, in contrast, tended to be black, chanting topless with their bodies covered in paint. Portrayed this way, they seem to revive a subversive strategy of destroying family values through moral and sexual promiscuity (Brito, 2020). Adherence to these demonstrations is of fundamental importance here too. Whereas the former was represented as amassing great numbers of demonstrators, the latter, so it was claimed, consistently failed to match them. In sum, changes in the red cluster indicate a persistent effort to delegitimize the president’s opponents.

This path of radicalization is not new. In diversifying the targets of their attacks, these far-right parliamentarians resorted to historical tropes which operate as euphemisms for alternative power relations. The discrediting of their opponents is grounded on the treatment of any political adversary as belonging to a “parasitic state preying on the people” (Carvalho, 1998, p. 250), a recurrent framing device of postcolonial societies associated with personalistic political loyalties. The representation of perilous oppositional protesters is also recognizable from a historical perspective. Chauí (2000) reminds us how the “Tradition, Family and Property” movement, which laid the ground for the 1964 military coup d’état, also accused leftist opponents of promoting the decay of Christian civilization and its values.

Green Cluster: Signifying nationalism

The green cluster consisted of visual messages mainly referring to political achievements. With a distinctive use of colors associated with Brazil’s national flag, the images exemplify how sensory patterns can be used to prime interpretative responses (Elliot & Maier, 2014), in this case, optimism with the Bolsonaro’s administration. These achievements are interpreted through nationalistic lenses as these parliamentarians claim that, to use one of the posts’ terms, the country “is on the right track” (see Figure 8, Appendix). The underlying sentiment evoked through these images is predominantly of pride in this political project. This pattern suggests a certain institutionalization (Laclau, 2005a) of right-wing populism in the first year of the legislature. That is to say, this period was marked by efforts to address the demands these politicians claimed to represent, tacitly accepting the democratic institutions they were elected for.

In its attempt to popularize Bolsonaro’s agenda, this expression of nationalism is but a new chapter in the long history of Latin American states making use of media institutions to legitimize their actions. Since the formation of the first populist governments in the 20th century, Brazilian political elites have used mass media to create national identities and to integrate the citizenry “as a bundle of individuals who partake in the same sentiment of ‘being Brazilian’” (Weffort, 2003, p. 40). Even if not unprecedented, it is important nevertheless to stress the extent that parliamentarians have become involved in the promotion of the government in today’s media landscape. Ultimately, this involvement has even raised questions about these parliamentarians’ ability to unreservedly express their opinions, as state agencies have been tasked with monitoring their activities on social media. ⁷

The green cluster presented significant changes, as illustrated in Figure 5. Its core space is now occupied not only by images convening the institutionalized nationalism described above but also by images calling upon Bolsonaro’s supporters to participate in massive demonstrations, particularly on the occasion of Independence Day in 2021. Against a backdrop of institutional instability, confrontations with the Supreme Court, and difficulties in approving bills in Congress, far-right populists began circulating visual messages asking supporters to manifest their approval of the government. With this intent, the 2021 green cluster conveyed the sentiment of a government under siege, which needs the prayers and continual demonstrations of popular support.

Figure 5.

Examples of images included in the green cluster of 2021.

Moreover, their visual messages nourish the belief that the government is on the side of the people, despite what the elites may say. The 2021 image mosaic contained numerous images “proving” the popularity of Bolsonaro’s government, supposedly countering the “narrative” from survey polls and mainstream media; in doing so, they asked their audience: who are they going to believe, the establishment or their own eyes? Interestingly, they also ridiculed conventional accounts that expressions of support for the president were nothing more than bot activity: “I have never seen so many bots!” was the text included in a photograph of a crowd (seemingly) expressing support for the president.

These images suggest a new way of signifying nationalism. We noted how the 2019 image mosaic revealed an institutionalization of this populists project, with the bulk of its posts seeking to disseminate (dis)information about the government’s achievements. Though this effort continued, we also began to observe, in subsequent years, a form of patriotism that summons “the people” not precisely to support the government’s agenda, but to save it from the imminent threat of its opponents. In this way, these frames are evoked in explicit confrontation with the “hypocrite,” “leftist” adversaries. And by giving “evidence” of popular support through these tweets, these parliamentarians reinforce the use of social media as an alternative to mainstream media institutions, thus furthering their populist logic.

Discussion

Populist communication has been faulted for being centered around the figure of the leader in detriment of bolstering the voices of their followers (Waisbord & Amado, 2017). This lack of voice reproduces a subaltern condition of democratic engagement (Santos, 2018), one in which discourses in the name of “the people” abound; however, little is known about the aspirations and living experiences of those who constitute that popular identity. Our findings suggest that this criticism can be extended to the visual communication of far-right populist parliamentarians, who play a crucial role in the formation and maintenance of populist projects.

The first two clusters identified, consisting predominantly of white and black colors, included images curating content from mainstream media, official documents, and social media accounts. These posts provided interpretations for Brazil’s state of affairs. Beyond their informative function, these clusters are also repositories of cultural values and attitudes. Their centrality in the image collection suggests an important development in the study of populism. If the pursuit of ideological dominance is characterized by the neutralization of contents which pose a threat to it (Laclau, 1978), here we see the extent to which far-right populist parliamentarians have become involved in ideological struggles. From news headlines to excerpts from the Bible, these politicians offered their audiences an alternative communicative source. Far-right parliamentarians benefit from a culture of iteration (Foster, 2022) insofar as their communication goes beyond discrete propaganda campaigns. As we documented in this paper, at stake is not just the dissemination of messages about a particular issue or event; but rather the maintenance of cultural interactions whereby political identities can be continuously (re)produced.

While the white cluster expanded from 2019 to 2021, the beige diminished. This cluster offered a glance into the political rituals in which these politicians are involved, with numerous photographs of ceremonies which attest to the importance of creating a threshold between the “real acts of individuals and the power of general will” (Bergem, 2018, p. 255). But we also noted how the beige cluster played with such a liminal zone: as populists in power, they present a pompous, distinctive kind of leadership; but as populist representatives, they disguise that same authority by staging an outsider, ordinary character.

The red cluster was, predictably, associated with attacks against the opposition. But the nature of these attacks changed significantly. Over the course of this legislature, the reddish images evolved from sporadic attacks, targeting primarily former president Lula, to denunciations of “hypocrite” narratives in media and state institutions. The targets of their attacks have dispersed. And instead of celebrations for the end of so-called leftist regimes, as observed in 2019, these images began symbolically investing the enemy with radical apprehension and threatening feelings.

The green cluster presented significant transformations too. With images depicting achievements of Bolsonaro’s government, this area initially exhibited as its main ideological function the absorption of “the people”/power contradictions within the system, what was previously described as institutionalized nationalism. This communicative effort persisted, but the next years introduced numerous images mobilizing supporters and, consequently, photographs of the mass demonstrations that followed. Consistent with the red cluster, these visual messages began nourishing a sentiment of government under siege in need of constant manifestation of popular support. Thus, patriotism ceased to be solely an expression of approval for the government’s agenda, as “the people” are now also summoned to save the presidency from the imminent threat posed by the enemy.

Insofar as these parliamentarians’ social media pages are concerned, president Bolsonaro’s conservative agenda enjoys great popularity, despite opinion polls claiming otherwise. Twisting Benkler’s (2006) original optimism with digital models of accreditation oriented toward “see for yourself,” far-right populists used their online accounts to “give evidence” of how widely approved the government is. They counter the “narrative” coming from surveys and mainstream media with the photographs they posted, telling their followers not to believe the mainstream narratives but their own eyes.

It is important to acknowledge the limits of this study. Notably, our visual analysis focused on Twitter. While Twitter is an essential social media platform for political actors, it is by no means the only platform they use. It is unclear whether these findings would replicate in other platforms, such as the image-based Instagram or the encrypted messaging services WhatsApp. For example, we could attribute the predominance of black-and-white images to Twitter’s position in our media ecology as a source of news (Fontes & Marques, 2022). However, there is also evidence that these results may replicate, as some politicians circulate near-identical content across platforms (Bossetta, 2018) adopting coordinated and complementary strategies within multiple social media (de Barros et al., 2021). Future cross-platform research could help us answer this question, but insofar as Twitter is concerned, the existing literature is clear in stating that this platform is not necessarily conducive to horizontal, bottom-up communicative interactions between citizens and populist leaders in Latin America (Waisbord & Amado, 2017).

Conclusion

What conclusions can we draw on how “the people” figure in these discursive formations? We noticed how the first clusters relied primarily on the content produced by the same institutions these parliamentarians antagonized with, thus failing to present us with visual or textual messages coming from their constituents. Next, representations of “the people” were also absent in their attacks against the opposition, as well as in liturgical or intimate occasions revealed by the beige cluster. As illustrated by Figure 10 (Appendix A), far-right populists depicted “the people” predominantly through mass gatherings, rendering tacit the voices of those who participate in them.

Represented this way, these constituents are subscribed to a quintessential form of a subaltern populism: they matter insofar as they deliver the unified, coherent, popular support needed for their leaders. Representatives claim to speak on their behalf, yet little is known about their supporters besides their adherence to such political project. This pattern is what confers a subaltern quality to their populist political communication, as conceptualized in the literature review. Of course, these protestors have chosen to be in those mass demonstrations. They do not lack a political voice altogether. However, this adherence should not prevent us from critically analyzing how parliamentarians frame their constituents, thereby filtering and giving salience to certain forms of political subjectivity. It is this process of making representations of “the people” which is being emphasized here. From this perspective, at the core of these parliamentarians’ imagery lies a homogeneous constituency whose contribution to their populist political project is difficult to grasp “beyond the raw material they represent” (Scott, 1998, p. 155).

In conclusion, while populism has been the subject of interest of many scholars, there is little literature exploring the use of visual content in furthering this phenomenon. In this article, we explored how far-right populist parliamentarians contributed to this political logic with the use of social media imagery. Even if populism had in the past a restorative role within democracy (Silva & Vieira, 2019), particularly in Latin America (Conniff, 1981), here we observed how parliamentarians’ attempt to (re)construct an ideological domain hindered efforts to signify and cultivate an authentic popular representation of “the people.” This constituency tended to be exclusively defined by the battles set in motion at the level of political affiliations. Indeed, all other dimensions of social reality seem to be organized and have their meaning derived from these antagonistic relations. The voices of those claimed to be represented by these parliamentarians were diluted in visuals of mass demonstrations, which provided the primary occasion for their incorporation into populist communication.

Supplemental Material

sj-docx-1-sms-10.1177_20563051231177962 – Supplemental material for Representing “The People”: What Can Social Media Images Reveal About Populist Propaganda in Brazil?

Supplemental material, sj-docx-1-sms-10.1177_20563051231177962 for Representing “The People”: What Can Social Media Images Reveal About Populist Propaganda in Brazil? by André K. Rodarte, Torie Hyunsik Kim and Josephine Lukito in Social Media + Society