RESISTING ALTERNATIVE IMAGES: An Ethnography of Visual Disinformation in Brazil

MIHAI ANDREI LEAHA

University of Barcelona

https://orcid.org/0000-0002-3684-9693

ROGER CANALS

University of Barcelona

https://orcid.org/0000-0002-1529-5943

---

November 2022, Rio de Janeiro: in her living-room-turned-office, the journalist Natalia Leal starts her daily routine. She is the CEO of Lupa, Brazil’s main independent fact-checking agency.¹ The night before, Luiz Inácio Lula da Silva had won the presidency in arguably the most disputed election in the history of Brazil. Natalia opens the Lupa website and begins to go through the first fact-checked images and video clips of the day. Brazil seems ablaze: video footage depicts armed drug gangs taking over cities (Figure 1);² videos show trucks blocking the highways; posts by “patriots” call for armed military intervention. Other images claim to depict the first signs of Lula closing the churches. Most of these images, Natalia tells Mihai,³ are photographic or imitate a photorealistic style. They are not necessarily “false,” in the sense that they are not artificially generated or intentionally manipulated to alter their initial composition. Nonetheless these images are deceptive when taken out of context or strategically associated with other images or written texts. “True images,” Natalia explains, “can end up conveying meaning divergent from the facts they allegedly portray.”

While Mihai films her, Natalia describes how deceptive images function as disinformation devices. Images spread fast and circulate in uncontrolled ways on social media, forming closed informational bubbles. Trust plays a key role in this process of viral disinformation: people tend to accept as “true” images coming from their “trust circles.”⁴ Natalia’s words point to an important issue: assessing an image of actuality as true depends on more than what the image shows. It depends, as we argue in this article, also (or mainly) on what we call the relational network in which it is inscribed.

Natalia next shows Mihai a video verified by Lupa that depicts the civil police storming the Brazilian Congress. The video circulated online as an account of a current fact. But the scene it shows corresponds to past events.⁵ Removed from its initial context, this video has become an instance of networked disinformation. It forms part of what Natalia calls “the main false narrative of the Brazilian elections of 2022: electoral fraud.”

These images seek to emotionally mobilize the audience to intervene in the world. As Natalia says: “They call for future actions.” They are thus performative, since they contribute to actualizing the reality they allegedly represent. Such images are less representations of the world than what we call “anticipatory devices” aimed at sparking political imagination and action.

Weeks after its publication (on January 8, 2022) the potentiality of this image was actualized in the biggest antidemocratic event in Brazil since the military coup of 1964: thousands of people invaded the Brazilian Congress (Figure 3) in a way that mirrored the United States Capitol attack on January 6, 2021 (Figure 2). “Images can be mobilized to anticipate the future by imitating or revising images from quite different social contexts and historical moments,” Natalia stated. Such images are usually ones people already know. So the general property of images becomes exploited by deceptive anticipatory images: images relate to reality, but they do so by relating also to other images. To quote W. J. T. Mitchell (2010, 15), the “war of images” is a war of cloning images.

This article offers an anthropological discussion of what we will call “alternative images.” We define alternative images as images with a deceptive referential value. While presented as accounts or reliable metaphors of reality, alternative images are in fact intentionally misleading. Alternative images do not need to be totally false (in the sense of manipulated, altered, or plainly “invented”) to convey a deceptive message. The falsity or deception of an image, we show in this article, is not a property of the image itself. Rather, the deception forms part of a relational, multimodal, and contextual network of images and messages. By using the term relational here we do not only point to the relationships that people maintain with images and through them but also, as the previous example shows, to the relationships that images maintain among themselves. These constitute multimodal relations. Multimodality here refers to the combination of a variety of communication elements such as texts, videos, audio messages, designs, emoticons, and GIFs, thus constituting semiotic wholes. Hence, alternative images must be analyzed in their misleading settings, specific media contexts, and channels of distribution (Chadwick and Stanyer 2022). Alternative images make for complex and intentionally distorted visual narratives with a performative potentiality that often plays a crucial role in disinformation and covert influence operations, two key concepts of our article worth defining now.

The concept of disinformation points to a set of deliberate untruths or misrepresentations, frequently disseminated as “reliable information,” spread with the aim of furthering political objectives such as undermining adversaries, disrupting policy discussions, swaying voters, exacerbating ongoing social tensions, or generating widespread confusion and information overload (Bennett and Livingston 2020). Disinformation is intentional and premeditated. It can have real and disastrous consequences for what we commonly understand as democratic elections.

Influence operations, on the other hand, form part of covert networked disinformation wherein the delineation of interests steering media content becomes obscured by instilling skepticism regarding the genuineness of the content and the motives underpinning its dissemination.⁶ Such influence operations do not necessarily make false claims, push propaganda, or launch attacks on adversaries. Rather, they present as multimedia content, familiar metanarratives of the kind often seen in disinformation campaigns (Gaw et al. 2023). Covert influence operations use alternative images to construct a “parallel reality” beyond the false/truth distinction in which a (false) sense of “reality” is enhanced by their dissemination and circulation. In these ways, alternative images acquire the power to influence common opinion.

This article has three main sections. In the first, we situate the concept of alternative images in the context of a more general debate on disinformation and “visual trust” (Canals 2020).⁷ The second section explores the online circulation and modulations of alternative images in the context of Brazilian electoral disinformation. Here we look at several case studies that account for the complex semantic and digital structures of alternative images. We also analyze their modes of circulation and manifestation in the digital landscape. In the third part, we adopt a praxeological approach to disinformation (Lelo 2024) so as to ethnographically analyze actions designed and undertaken to face the political influence of alternative images. The specificity of these actions—which vary from carnivalesque embodied manifestations to street interventions and visual verification techniques—lies on the fact that they react to alternative images either by creating and mobilizing new images or by implementing visual technologies aimed at assessing the reliability of images. In other words, they fight images through images. We propose to define these actions as modes of “iconic resistance.”

PICTURING ALTERNATIVE IMAGES AND DISINFORMATION

The rise of fact-checking agencies like Lupa needs to be framed within a tumultuous present in which natural disasters, wars, and pandemics coexist with a crisis of information described by authors like Victoria Rubin (2022) as “infodemics.”⁸ Basically, this concept refers to a new post-truth mediascape (Barclay 2021) in which distorted facts (and images) can produce severe informational disorder that affects trust in institutions and mass media, threatening democracy, freedoms, and basic human rights (Filimowicz 2022; Grech 2021). As the infamous case of Cambridge Analytica illustrates, science and technology can be used to influence behaviors, change minds, and manipulate the audience politically (Kaiser 2019). In addition, the advent of generative multimodal artificial intelligence (AI) and the use of algorithms in decision-making processes is redefining politics and visual communication as we know it (Bucknall and Dori-Hacohen 2022; Bucher 2018; Coeckelbergh 2020). In this context, fact-checking agencies portray themselves as rigorous, objective, and transparent platforms doing a public service of media literacy. They aim to assess the flux of information in different media to clearly distinguish objective facts from distorted and false narratives. Among these false narratives we find what we call “alternative images.”

The term alternative images as we use it here paraphrases the expression “alternative facts” coined in 2017 by Kellyanne Conway, an advisor to former U.S. president Trump. Conway came up with this term when defending attendance figures at Trump’s presidential inauguration. The disagreement that led to the term alternative facts related to how the event was photographed. Conway accused the journalists who made these images of having taken them from a biased, bird’s-eye view. According to her, the blank spaces in the images of the crowd resulted in a distorted depiction of the large numbers attending the event. Despite the facts showing a much larger crowd at Obama’s inauguration (Fandos 2017), Conway defended the assertion that Trump boasted the highest number of attendees in history. “And that is an alternative fact,” she stated. According to Conway’s logic, no images are more “accountable” or “truer” than others. The veracity of what an image portrays would be more a matter of belief than a matter of facts.⁹

Conway’s “theory” had a big impact in American society. The more it circulated, the more the facts of the presidential inauguration became distorted. As Barclay (2021) keenly observed, disinformation spreads in the same manner as folk songs: the oral history–like transmission of disinformation narratives tends to enrich meanings, forget origins, and add new senses to the initial message. Much like Richard M. Dorson’s (1950) concept of fakelore, alternative images and disinformation narratives rely on the “invention” of new meanings rather than on simple repetition or on the principle of “visual traceability” (Canals 2020).

As Selma Ottonicar (2020) argues, users actively interact with and get attached to images from known or trusted networks. They often end up paying the images forward, potentially making them go viral. Trust thus emerges as a strong principle in these processes of image-intensification (Canals 2020). The concept of image-intensification refers to the enhancement of the potential meanings and functions of images. It is through acts of intensification that images acquire new significance and roles, thus recursively affecting the context in which they find themselves. The importance of creative circulation and trust in the study of current images show that images should not be studied from a static and abstract point of view, but, rather, from a dynamic and relational perspective (Freedberg 1989), focusing on their “iconic paths” (Canals 2020) and analyzing ethnographically the relationships individuals establish with and through them.¹⁰

The term alternative as we use it in the concept of alternative images comes from the same root as altered. Both concepts point to the idea of changing the status or nature of someone or something. In this sense, we will focus in this article on how alternative images alter reality, thus having a performative nature. Yet our objective is not to emphasize (once again) the alleged power of misleading media content on the general population. We feel uncomfortable with the already deconstructed idea of “powerful media and weak audiences” (Livingstone 2019). Images do not do things only by themselves. For this reason, inspired by a communicative and imaginative approach to disinformation (Cabañes 2020; Andersen and Søe 2020), we highlight the role of human agency not only in the creation and circulation of alternative images but also—and especially—in the strategies carried out to counter them visually.

From a methodological point of view, our research is based on a multimodal ethnography carried out by Mihai during the 2022 Brazilian presidential election, a moment of deep iconic intensification (Canals 2020). Mihai conducted the ethnography in both online and offline environments, using collaboration, experimentation, intervention, and invention strategies to account for complex media and political ecologies (Collins, Durington, and Gill 2017; Dattatreyan and Marrero-Guillamón 2019). The online/digital research was carried out by observing the circulation of alternative images on social media such as X, Instagram, and Facebook, but also in several WhatsApp and Telegram groups. Moreover, the network and accounts of politicians as well as of campaign managers and popular political influencers were followed on main Brazilian TV channels, YouTube, and podcasts. A special focus was given to the right-wing politicians and influencers, as they emerged most frequently as the agents of disinformation during the 2023 campaign.¹¹ We were particularly interested in how these media became distributed, enforced arguments, and instrumentalized images for their political purposes. Images that would support or help disinformative arguments were then archived for further analysis. We followed the movement of online images from the perspectives of visual analysis and visual communication (Ledin and Machin 2018; Aiello 2019). To fully understand the context of networked disinformation in Brazil, we also tracked images back from the verification divisions of media outlets such as Folha, UOL, or Globo, among others. We surveyed investigative and political analyses on podcasts and journal articles on Nexo and Piaui. With this approach, we managed to follow the “iconic path” of digital images through fact-checked images debunked by independent agencies such as Lupa and Aos Fatos, or by the fact-checking divisions of media trusts such as Estadão Verifica and UOL Confere.¹²

We used ethnographic cinema to depict some of the above-mentioned movements of iconic resistance. On the streets of São Paulo, Mihai filmed and collaborated with left-wing artivists engaged in visual street interventions as a response to the disinformation campaign of Jair Bolsonaro. He also filmed, through both an observational and participatory approach, different actions carried out on the streets by various political supporters (projectionists, supporters, activists, and performers) and gatherings in squares or central areas of Rio de Janeiro and São Paulo. We included campaign images and poster bombs in our ethnographic footage, since we considered them agents of iconic resistance themselves. The footage that resulted served both as primary ethnographic source for writing this article and as the basis for creating a short research film titled Verdade e Mentiras (Truth and Lies, 2023) that we will describe and share in the third section. Doing research with a camera permitted us to account for the sensory and performative character of political events that goes beyond the textual ethnographic description. This research follows a long tradition in visual anthropology that aims at studying images through images (Canals 2023).

Despite the recent and rich amount of literature on post-truth, fake news, and disinformation, not much has been said about the role of images in disinformation strategies. Less has been done in terms of ethnographies of “false images” adopting a bottom-up, relational, and qualitative approach. This is where our work intervenes from the following section onward in relation to the case of Brazilian elections.

ONLINE MODULATIONS OF ALTERNATIVE IMAGES IN BRAZIL

In Brazil, disinformation is a hot topic. As we are writing this article (September 2024), draft law 2630—or the law project (PL) of “fake news”—is still being discussed in Brasilia. The project wishes to regulate online disinformation and impose obligations on the big tech platforms that host the dissemination of disinformation. The project aims to tackle, not without controversy, the chaotic context of disinformation that has taken over in Brazil in recent years. This context started to take shape in 2017 alongside the rise of the far right and the increased popularity of the charismatic figure of Jair Bolsonaro (Gouvêa and Castelo Branco 2022; Lapper 2021; Klem et al. 2020). The journalist Patricia de Campos Mello (2020, 74) ascribes Bolsonaro’s rise to power to a larger context of global “techno-populism” (Bickerton and Accetti 2021; Da Empoli 2019). Campos Mello (2020, 10–37) argues that Bolsonaro took power in 2018 with the help of a “digital militia”—or a “hate machine,” as the journalist calls it. The movement was articulated mostly online and relied heavily on the propagation of conservative political values. The covert influence operation involved, among other elements, the discreditation of mainstream media and journalism and the use of hate speech. The scheme relied on a massive spread of disinformation through WhatsApp messages and social media. According to Campos Mello, this strategy helped Bolsonaro win the election in 2018.

In 2022, it became known that Bolsonaro’s campaign would be based heavily on networked disinformation and that most of the false and misleading content was to be targeted at his opponent, the former president Lula. The investigative journalist Breno Pires explains how the Workers’ Party (PT) reacted to the right-wing digital militia led by Carlos Bolsonaro (Pires 2022).¹³ In the February 2022, the sociologists Fernanda Sarkis and Marcus Nogueira formed a team with IT specialists, marketing experts, and lawyers hired by the PT to help deconstruct the right-wing disinformation network (Figure 4). They based their methodology on the concept of a “cartography of controversy,” inspired by Bruno Latour’s Actor-Network theory, according to which, to understand any social actor, you first need to understand their behavior within the network. They discovered that “Jair’s show”—as the two researchers called it paraphrasing the film The Truman Show (Weir 1998)—produced a media ecosystem of disinformation managed by eighty-one profiles involving three of Bolsonaro’s sons and many other high-profile parliament members, bloggers, YouTubers, and influencers.

According to the study, the thematic threads of networked disinformation were concentrated around four main axes: criminality and violence; religion and customs; electoral system credibility; and the socioeconomic agenda. The false narratives used images to show associations between criminal groups, Lula, and the PT; false depictions of a non-Christian, communist Lula; images portraying Lula as Lucifer in an attempt to create moral panic among Christian voters; deep-fake videos falsifying electoral research; decontextualized images depicting electoral fraud; and images of political rallies that were either manipulated or out of context, to name just a few.

In the following, we are going to analyze how right-wing networked disinformation used alternative images as they appeared in our online research. Drawing on the work of François Laplantine (2015), we use the term modulation to refer to a double movement of alternative images: the movement of images “being altered” by actual events (and having an influence on them), as well as the one of images being altered by users in an endless process of copying and differentiation. The former has to do with the principle of referentiality (how images relate to external reality); the second with the one of intensification (how they are transformed when put in circulation).

Bolsonaro’s presidency was marked by frequent attacks on and the cutting of state funding for Brazil’s main media trusts. It constituted a coordinated, long-term attempt to discredit the free press and its institutes (Campos Mello 2020). For instance, Folha de São Paulo (Brazil’s most prestigious newspaper) was often accused by Bolsonaro of misrepresenting him. Folha’s poll research institute, DataFolha, became one of the victims of Bolsonaro’s attacks during the electoral campaign. For the right wing, the “misleading” information from DataFolha needed to be replaced with what Bolsonaro called Data Povo (Data People). Data Povo was an expression, a hashtag, and an image used to replace graphic scientific abstractions, such as official election-poll data visualizations.

Our online research consisted of observing on social networks how the profiles of top right-wing politicians, including Bolsonaro’s sons, articulated disinformative narratives and alternative images trough their network of influencers. For instance, during the 2022 electoral campaign, Brazil witnessed an unprecedented dispute about the “image of numbers,” as shown in election polls and beyond. One such example was what is known as “the image of the crowd.” On September 7, 2022, Bolsonaro invited his supporters in Brasilia to celebrate 200 years of Brazil’s independence. As a national celebration, the events were not supposed to be politically biased, and Bolsonaro was not allowed to make campaign declarations during them. Yet the event transformed into a right-wing demonstration of force, and Bolsonaro used the occasion for his political propaganda. As a consequence, the images of the September 7 event were banned from online circulation by an order from the Electoral Supreme Court (TSE).¹⁴ The images were removed from the main social media networks shortly after the event, yet Bolsonaro’s propaganda machine found a way to subvert the TSE’s decision. In an X post from September 8, Carlos Bolsonaro posted a photo of a huge crowd wearing the yellow-and-green national football shirts from the event in Brasilia (Figure 5).

The text accompanying the post reads: “According to DataFolha this is a sunflower plantation” (@CarlosBolsonaro, September 8, 2022).¹⁵ It echoed a declaration from Jair Bolsonaro’s September 7 speech saying, “I have never seen such a huge sea [of people] in yellow and green. Here there are no liars from DataFolha. Here is Data Povo [Data People]. This is the truth, and the will of honest, free, and working people” (Delgado 2022).

In an interview about the September 7 events in Nexo, the political scientist Jorge Chaloub observes that the events meant to legitimize the discourse of fraud against the research pools and electoral system, proving that they do not represent the real will of the people. It further advances the conspiracy idea that public and private institutions oppose the right-wing movement (Cruz 2022). The image of “the sunflower field,” we might add, subverts the image of scientific data with the image of an unquantifiable, “infinite” number of people. This shows that alternative images often constitute subtle ways of “accounting” for alternative facts within the context of complex influence operations. They can be “authentic images,” with a valid, actual referent to a historical moment, that are nevertheless used to deceive and support a disinformative discourse.

The film director Miguel Antunes Ramos (2023) observes how right-wing propaganda carefully forged the image of the “Bolsonarist” crowd. These images often contain an amateurish framing that produces a view of the crowd created by “anonymous authors.” Bolsonaro’s propaganda machine produces images for and through his base. The view of the crowd is validated through an aesthetic of authenticity that is recognizable, likeable, and sharable. The perspective of the crowd becomes an alternative perspective turned into a political trophy, ready to be taken home and archived (McKay 2020). It is a bottom-up propagandistic approach in which unknown authorship and “viral sharing” have the power to replace scientific data visualizations with politically effective alternative images.

Networked disinformation uses alternative images as misleading social decoupages. Like the image of the infinite crowd, visually manipulative techniques such as scaling, reframing, or cropping can distort the “bigger picture,” leading to misrepresentations of the social context. For instance, on September 1, 2022, the self-declared conservative newspaper Revista OESTE published an article titled “Indigenous People, Supporters of Bolsonaro Organize an “Indio-ciata” (Redação Oeste 2022).¹⁶ The paper used the following image as its cover:

The image is a frame from a video showing what appears to be an Indigenous event in support of Bolsonaro. The article argues that the event proves support from some Indigenous groups for mining in Indigenous lands. Flavio and Carlos Bolsonaro posted the video of the event on X, insisting on the images of the Indigenous group carrying the flag. On X, Flavio Bolsonaro shared the image, adding the following text: “Our Indigenous people are (with) Bolsonaro.” A day later, a report by UOL disclosed a very different story (Audi 2022). The “bigger picture” reveals that the video depicts a sacred Xingu ceremony, called Kuarup. The group performed this ritual in Xingu Parque to pay homage to the victims of COVID-19. The Indigenous local leader and activist Watatakalu Yawalapiti explained for the report that the two Indigenous people carrying the flag were not from their village and that they were given alcoholic beverages to bring the poster to the ritual. She also declared that only a small minority in the Xingu territory supported Bolsonaro’s re-election. During the dance, Watatakalu had banned any political manifestations, considering it disrespectful to the victims of COVID and the ritual. She believes that the former government’s delay in giving the community vaccines was partially responsible for the deaths of her kin. In addition, in the following days, the Association of Xingu Indigenous Land (Associação Terra Indígena Xingu, or ATIX) launched legal action asking the Supreme Electoral Court, the TSE, to remove the misleading tweets.

Other, more provocative alternative images and messages circulated online and offline during the 2022 campaign, as did memes. Memes have gained importance in the digital political sphere, leading Ricardo González-García (2019) to coin the term memecracy. Creating memes is a form of visual creativity that may include parodic imitation and ironic camouflage (Shifman 2014; Kien 2019). Networked disinformation incorporated memes to emphasize the right-wing narratives and make them more effective. Our research revealed many right-wing memes circulating online. Indeed, as the Brazilian philosopher Paulo Ghiraldelli (2019, 104) noted, “nobody else has produced as many memes as Bolsonaro in [Brazilian] politics.” Like caricatures, memes use exaggerations and add impact phrases to convey a political message.

In this meme circulating among far-right groups, Bolsonaro appears like the savior of Brazil, portrayed as a Rambo figure fighting corruption. It reflects an ideology that supports the wide availability of weapons and military power. The image hints at what João Felipe Gonçalves and Gideon Lasco (2023, 59) call the “nefarious affinities” between economic neoliberalism and political illiberalism by picturing Bolsonaro as a “neoliberal populist,” that is, as a avenger supported by the masses. The meme also conveys the aggressive masculine character common in contemporary populist performances (Aciksoz and Korkman 2017). This alternative memetic image contributes to the construction of a parallel reality of a Brazil at war and needing to be rescued by a charismatic, violent, and macho hero.

But images can hit back, and anti-memes can react to political leaders’ outrageous statements using humor or even black humor to convey a subversive political message (as we pointed out at the beginning of this text: images relate to things, but they always relate to other images too).

This neo-protestant iconography depicting Jesus and Maria Magdalene was posted on social networks and is accompanied by the following text: “I don’t rape you because you are not worth it. Signed: Messiah, Jair (Bolsonaro) 2014.” The text quotes a statement Bolsonaro made in 2014 to the congresswoman Maria do Rosário. The meme not only refers to that particular moment but also attempts to subvert the image of Bolsonaro as a Christian athlete with a contradictory analogy between the image of Jesus and the politician’s misogynist statements. In Brazil, religion as an electoral theme has seen a rise in popularity since 2018. According to DataFolha research from August 2022, 49 percent of Brazilians considered the subject of religion very important in the 2022 elections (Redação Folha de S. Paulo 2022). Bolsonaro has built strong support among evangelical voters, and research points to a connection between the growing popularity of an evangelical and the conservative movement in Brazil (Spyer 2020; Araújo 2019; Alexandre 2020). The war of memes and counter-memes shows that memes can play the role of political radar, which has the power to mobilize the population (especially young generations). They may be designed to rail against an illiberal regime (Kallius and Adriaans 2022) or to reinforce conservative positions.

Yet memes materialize beyond the digital sphere (Pink, Ardèvol Piera, and Lanzeni 2016). While filming with the artist Julio Djocsar, we encountered the following poster bomb on the streets of São Paulo:

In this poster bomb, quotes from Bolsonaro, including the one about the rape, accompany a deformed portrait of the politician wearing a sheep mask. The Bible quote reads like this: “Beware of false prophets, which come to you in sheep’s clothing but inwardly are raving wolves.” This image blurs the distinction between political and religious images, and shows how intensification, as the creative movement of images from one support to another, can be used to fight false narratives. We will delve deeply into the question of iconic resistance in the following section.

REACTING TO AND RESISTING ALTERNATIVE IMAGES

Alternative images and disinformation may contribute to producing different political effects and emotional reactions. For those who “trust” them, alternative images offer fair accounts of or metaphors for reality. For those who do not trust them, alternative images are simply visual lies that can provoke indignation, rage, and a general feeling of injustice. These feelings may lead to actions of iconic resistance spreading across offline and online environments. In this section, we tackle some of these “iconic paths” by ethnographically depicting how various social agents mobilize, multiply, react to, and resist alternative images on the streets and beyond. One of the specificities of these movements is that they are not iconophobic: they fight “visual falsity” through images, thus creating visual interventions (Pink 2007) with the purpose of impacting the public sphere, thereby blurring the offline-online distinction (Favero 2018).

Performing Images in the Streets

The first instance of offline iconic resistance to alternative images that we analyze relates to the use of the body as a public image in demonstrations, especially against Bolsonaro. The demonstrations in support of Lula and against the spreading of alternative images took place in a general context of emotional intensity. This connects to one of the senses of the term alternative images as we have defined it: these are images that have the power to “alter” people.

In Brazil, due to the electronic voting system, election results are known just a few hours after voting has finished. In 2022, Lula was announced as the winner in a very close presidential race. Young, middle-class supporters celebrating his victory in Rio’s São Salvador Square burst into tears. The crowd was euphoric, “altered” or “possessed” by the spirit of national victory. Political and communication studies rarely consider the agency of the body as an image in a political dispute. Taking into account what Laplantine (2015, 65) calls “an anthropology of the sensible,” we consider social facts as sensorially determined and emotionally performed. In this sense, the role of the individual and the agency of the body become politically relevant.

During the campaigning period, images of Lula and Bolsonaro were exhibited, shown, and performed by supporters in a kaleidoscope of embodied and politically symbolic gestures. On the streets, bodies were performing images as much as people were “being performed” by images. For example, at a weekly neighborhood fair in Rio de Janeiro, a life-sized image of Lula “came to life” as passersby hugged, talked, and took photos with the image as if it were real, in a clear instance of what we call “iconic intensification.”

We agree with Tancons (2015, 136) in the idea that carnivalesque manifestations in political manifestations can be read as “organizational devices for the creation of new subjectivities prefigurative of participatory democracy through transformational processes.” In Campinas (São Paulo state), left-wing supporters decided to organize a carnival event for Lula named CarnaLula (Figure 11). Various carnival blocs—mostly white middle class from Barão Geraldo, Campinas, like Berra Vaca, Caxeirosas, União Altaneira, Cupinzeiro—participated in the event. One of the organizers explained that they wanted to show to the inhabitants of a peripheral and politically undecided neighbourhood in Campinas that Lula supporters also defended national values and identity, and that they could act peacefully and joyfully. This was a reaction to alternative images depicting Lula as an anti-patriotic figure. The group was also fighting one of the most important disinformation narratives of the campaign, which depicts Lula as a corrupt thief connected to crime. The carnival blocks were dancing and playing traditional carnival songs, dressed in red and showing Lula’s image either on T-shirts, stickers, flags, or drums. Lula’s images became modulated though dancing and performing. The bodies of the dancers as well as the voices of the singers turned into agents for performing political images. The agency of the body made the images “come alive,” while the images “modulated” the body and made it a device for political manifestation. The images of the event were then uploaded online with the hashtag “Lula love.” The iconic path of these images continued in the online sphere.

Iconic Modulations as Interventions

Across Brazil, groups of visual artists resist disinformation and alternative images in highly creative ways. During the electoral campaign, we invited the São Paulo–based visual arts collective Arte em Fluxo to participate in a multimodal experimental collaboration. It was an inventive attempt to combine ethnographic filmmaking and artistic interventions. We had previously discussed with them the possibilities of organizing a visual street intervention in which the artists Julio Djocsar and Cauê Maia would respond to the avalanche of fake news circulating in Brazil. The process would then be filmed and included in a short ethnographic film. The experiment featured two street interventions. The first one responded to Bolsonaro’s questioning of a study by the Food and Agriculture Organisation of the United Nations, which revealed that Brazil had once again entered the world hunger map as of July 2022. Bolsonaro had attempted to discredit the finding, saying that he did not see any people begging for bread in front of bakeries. The artists felt compelled to respond to the spread of these false statements. Julio, a long-term practitioner of poster bombs, had the idea of creating a large lambe-lambe (the Portuguese name for poster bombs) measuring 2.5 meters, featuring the word VERDADE (truth). The intervention consisted of strategically placing these posters under bridges, in parks, and on walls in the city center, places frequently passed by individuals without stable housing. They were the first ones to be impacted by the intervention, and some of them commented on camera about the importance of establishing the “truth” about Brazil’s severe social inequalities. The filmed intervention aimed to re-establish the significance of the word truth by linking it to the immediate and visible reality of São Paulo.

In their second intervention, the artists shifted their focus to the antonym of truth: MENTIRAS (lies). Cauê Maia forms part of the Transverso Collective,¹⁷ and he creatively uses the sprawling urban environment of the South American metropolis as a canvas for his politically and poetically charged artistic interventions. Enabled by a portable and adaptable projecting device called “bazooka poetica,” designed by Cauê (Baptista 2019), his captivating projections occupy the blank walls of São Paulo.¹⁸ The bazooka consists of a lantern and a magnifying glass. It uses slide-like images printed on transparent plastic paper to negatively project on any surface. The user of the bazooka then passes through a process of adaptation, reframing, and fitting that aims to fix the urban poetry on the chosen surface. Through his poetry, Cauê subverts political messages and turns them into haiku-like short sentences. During the nighttime intervention, he and his collaborators projected poems and images onto various available surfaces of city buildings and towering skyscrapers in São Paulo. Their thought-provoking political messages serve as a resilient response to the prevalence of far-right ideologies and often address issues of social injustice, feminism, and Indigenous rights. The intervention was carried out on top of the Minhocão highway/park in the center of São Paulo. One specific image included the old blue bird Twitter logo and the following text: “The lie has short legs but knows how to fly high” (Figure 12). The intervention was this time a subversive form of denouncing Bolsonaro’s networked disinformation. The projections took images closer to city residents’ windows and then uploaded them to social media. The impact of the intervention went beyond the outdoor modulations into the online realm. The images the collaborators made during the intervention potentially impacted the 16,000 followers of their Instagram account.

The short ethnographic film Verdade and Mentiras (Truth and Lies) accompanying this text complements our ethnography as part of the visual depiction of “the sensible” in Laplantine’s sense. The multisensorial qualities of film, together with its ability to depict and creatively replicate rhythm, movement, and the corporeal, can account for the modulation of images in the two artistic interventions. Modal visual anthropology can be understood as an image-making process that implies both collaborative intervention and anthropological knowledge-making.

Iconic Assessments and Alternative Images

Alternative images circulate both online and offline. They spread so quickly that their origin and “iconic path” are hard to identify and almost impossible to trace. The rise of fact-checking and disinformation journalism has led to attempts to demystify the origins and trajectories of digital images through mixed digital forensics techniques (Graves 2016; Gogolin 2021; Sencar, Verdoliva, and Memon 2022). In Brazil, the fact-checking movement is recent and still struggling between economic sustainability and editorial independence (Lelo 2022). These agencies form part of international networks that support their “journalistic neutrality.”¹⁹

During the last three weeks of the October 2022 campaign, Mihai filmed and observed the work of Lupa agency. Since the pandemic, the team of Lupa works from home and connects, organizes, and publishes their articles online. The members of the team accepted our invitation to spend the most intense days of the campaign with them and accompany their coverage of the campaign. We were particularly interested in understanding how fact-checkers assess, verify, and are affected by alternative images.

Chico Mares, one of Lupa’s fact-checking specialists, explains how the agency assesses allegedly false images. They use specific digital software for operating reverse internet search tools while trying to determine the online iconic paths of the verified image. He claims that once an image is circulating online, it is almost impossible to identify the original. When performing so-called visual forensics, fact-checkers can feel frustrated before the impossibility of objectively assessing images. Nevertheless, looking for past versions of the same image often proves useful. To do so, Mares searches for previous publication dates and contexts and then attempts to establish a chronological circulation path for the image. In this way, fact-checkers can determine whether an image was taken out of context and re-mashed, modified, or re-signified to fit disinformative purposes. These techniques prove useful when tackling alternative images created through direct image-editing or montages, because the initial image allows the verifier to assess the changes in the image’s iconic path. Mares indicates, however, that images do not allow for as direct a verification process as textual information, mainly because they are harder to label as false/true or manipulated. “Images are not facts,” he states.

Images take time to fact-check, and by the time it is done, the harm of disinformation may already have been done. This causes frustration among fact-checkers, who find themselves in a race against time. The fast-moving circulation of alternative images in closed information bubbles makes the verification process even harder. Moreover, the Lupa fact-checkers mention that alternative images can exert a kind of violence on journalists—they emotionally alter them. Images can contain disturbing graphic content that might expose fact-checkers to shocking depictions. Fact-checkers also often receive threats from people who consider their work biased and feel they have a hidden agenda against their favorite candidate. As a countermeasure to this violence, Lupa offers its employees psychological support, especially during electoral campaigns.

The main political, ethical, and methodological tenets of Lupa in understanding and countering the affects and effects of visual disinformation in a fast-moving and unregulated media ecology align with the thesis presented by Kimiz Dalkir and Rebecca Katz in their 2020 work Navigating Fake News, Alternative Facts, and Misinformation in a Post-Truth World. The first principle consists in fighting technology with technology, mainly by developing innovative software capable of tracking down and revealing online visual disinformation. The second principle concerns the importance of media literacy. In this regard, after the elections, Lupa shifted its main focus from fact-checking to disinformation journalism, with the declared intention of educating the public about how disinformation works. Lupa produces almost daily videos and photographic materials for a larger audience, as one can see on their website, where each piece of content is designed to fit specific social networks and audiences. In 2023, the agency invested in media literacy programs and helped create the first university master’s program in journalism and fact-checking in Brazil. “This is a long-term commitment,” says Natalia. “And a hard one,” she adds, “which needs institutional support.” At a conference on digital journalism in Huesca, Spain, in 2023, the Brazilian journalist also advocated for both new laws and state regulations to ensure the accountability of social media platforms that host disinformation.

Contrary to the studies that see fact-checkers as part of an iconoclastic movement (Andersen and Søe 2020), we interpret them as non-iconoclastic, since they do not deny the capacity of images to become reliable sources of information. They “just” want to debunk false and misleading ones. Actually, fact-checkers are active image-makers, since the act of labeling images (adding to them tags like “false,” “true,” or “deceptive) in itself constitutes an act of visual communication.

Despite its efforts, Lupa felt disappointed with the limited effects of fact-checking on voters. Many known political influencers and agents of Bolsonaro’s networked disinformation group managed to secure a place in the current parliament. Natalia advocates for new solutions to tackle disinformation. According to her, updating legislation and state regulations on social networks should be complemented by pedagogical programs and explicit support of fact-checking agencies and disinformation journalism. The first steps in this direction have already been taken: the Brazilian law PL2630 is trying to increase the responsibility of the platforms and update the laws on digital information.²⁰ Right-wing critics have deemed this law a new form of censorship that may threaten fundamental human rights such as freedom of expression.

As we write this article, Bolsonaro has been declared ineligible for office for eight years by the Electoral Supreme Court in Brazil (Mendes and Cury 2023). The main reason is the propagation of fake news.

CONCLUSIONS

In this article, we have defined alternative images as intentionally disinformative visual content aimed at creating a parallel reality. We have delved into the context of the Brazilian presidential elections by discussing how alternative images contributed to altering the Brazilian electoral mediascape. We have focused on two aspects: first, through online research, we analyzed the production and circulation of alternative images within the frame of a broader disinformative media ecology; second, through offline visual ethnography, we discussed different reactions to alternative images: from bodily demonstrations in the streets to fact-checking agencies or artistic interventions.²¹

We argue that the question of alternative images is intertwined with the notion of “visual (dis)trust.” For images, trust is a relational category (Corsín Jiménez 2011), a matter of “ordinary ethics” (Lambek et al. 2015). In everyday life, the truth stance of images is validated (or refused) within trust circles rather than through factual verification. Our online and offline research revealed that alternative images play a crucial role in building trust and consolidating disinformative narratives by creating communities of disinformation. Political influencers as agents of covert influence operations play a critical role in building and sustaining these trust networks, which are based more on emotional affinities than on political rational debate.

Covert networked disinformation often works through micro-targeting, with the help of influencers that can get the messages more precisely to the trust circle. In this sense, the sender, as part of a previous network or community of trust, endows the image with legitimacy through the act of sharing it (this is an instance of what we called “intensification”). Once someone trusts an alternative image as we have defined it, the image for this person ceases to be alternative—it becomes “credible” or simply “true.” In a nutshell: disinformation needs trust conditions and specific networks to function.

On top of this, trust is a future-oriented concept. The fact of trusting others (or trusting images) prompts us to do things. Trust therefore has a performative nature (Broch-Due and Ystanes 2016), as the invasion of Congress described at the beginning of this article shows: the people who trusted the image of the invasion of the Brazilian parliament ended up invading it. Yet this performative nature can be used as an anticipatory device to foresee future events and fight against the effects of disinformation. This is also what fact-checkers do. They know that when an alternative image circulates uncontrollably, it will likely produce effects and turn into reality (as Natalia prophetically announced when she commented on the images of the above-mentioned assault before it occurred). More empirical research is needed to fully understand the impact of alternative images within often non-homogenous “trust-circles” and among consumers of disinformative content.²²

The aesthetics of alternative images also play an important role in the building of trust as common imaginary. The meme of Bolsonaro as Rambo, for instance, inscribes itself into a specific conservative aesthetic involving a particular rhetoric of disinformation. Judging it strictly from a visual point of view, we might think of the image as an anti-meme, in other words, as a leftist joke working against Bolsonaro—but it is the opposite. This proves that images cannot be studied in the abstract, as independent visual signs. Their understanding must be rooted in complex and dynamic relationships from which they acquire an original meaning and may trigger specific actions.

When analyzing alternative images, we need to focus our attention on what we have called “movements of iconic resistance.” We have analyzed three ways of reacting to alternative images: embodied performances, artistic interventions, and visual verification. They all respond to a general feeling of distrust about images of actuality circulating in the current Brazilian mediascape (and elsewhere). Visual distrust in fact-checkers is mostly based on a kind of epistemological skepticism (they assume that, a priori, no image is totally reliable before a scientific verification). In this sense, it differs from the initial distrust that fuels the work of artists or carnivalesque public demonstrations, more connected to an action-oriented emotional and political engagement. The three instances of iconic resistance discussed also show differences in the way they operate. Artistic interventions or public demonstrations are not interested in proving that alternative images are “false” or “misleading”; instead, they wish to subvert them through public and creative acts of image-making and image-intensification.

From a methodological perspective, we acknowledge the necessity of addressing disinformation and alternative images not only from a communicational or semiotic point of view but also through a bottom-up and qualitative approach, combining different modes of doing research. This approach dissolves online/­offline distinctions and focuses on the paths and agency of images in society. We advocate for the importance of setting up ethnographies of disinformation aimed at understanding how alternative images affect and provoke reactions for specific people in concrete historical and sociocultural contexts. We also believe it necessary to study the role of images by using hybrid research and participative methods, including visual and multimodal ones.

Lastly, we think that this research on alternative images may recursively prompt us to reassess the responsibilities and potentialities of anthropology in post-truth cultures and disinformative environments. Do we anthropologists have a responsibility to position ourselves against disinformation? What is the reflexive and ethical stance that we should take before the proliferation of alternative images? As we have seen, within the debate about fake news and disinformation, there is a clear place for journalists, militant artists, and active citizens. Yet what would mark anthropology’s place?

We have no doubt that networked disinformation constitutes a threat to democracy, knowledge, and equality. We also assume that our work consists of building ethical relationships and collaborations with our interlocutors through which we provide reliable descriptions of reality, which in turn lead us to propose new ways of interpreting and imagining it. Like journalists, we anthropologists are committed to trust and truth. And like artists, our practice involves creativity, imagination, and critical thinking. So, how can anthropology help deconstruct disinformation through alternative images?

We consider that a first step in this direction is to foster what we have called “ethnographies of disinformation.” This kind of research may contribute to a better understanding of how alternative images work—and therefore how to debunk them. We also have to commit ourselves to the education about and development of visual literacy, along the lines of multimodal and public anthropology. Our research must manage to reach different audiences and have a voice in public debates. We must assume the risks of taking our knowledge beyond academia—to collaborate with artists, social movements, and fact-checking agencies.

The anthropological study of disinformation cannot be detached from the act of actively counter-fighting it through critical, rigorous, and engaged research.

ABSTRACT

NOTES

Acknowledgments We thank all the individuals and institutions that have contributed to this research, especially the fact-checking agency Lupa and Natalia Leal. We also give a special thanks to Transverso Collective and Cauê Maia for his support and cooperation both in Brazil and in Barcelona. We are also grateful to the editors of Cultural Anthropology and to the reviewers of this article, whose comments and insights allowed us to improve it substantially. This article is part of the ERC project Visual Trust: Reliability, Accountability, and Forgery in Scientific, Religious, and Social Images (PI: Roger Canals, 2021–2026; www.visualtrust.ub.edu). This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 Research and Innovation Program (grant agreement No 101002897).