Social Representations of Political Polarization through Traditional Media: A Study of the Brazilian Case between 2015 and 2019

1.1 Media and SRs

Social representations (SRs) are part of our everyday thinking, giving coherence to our ideological beliefs (Moscovici, 1988). Through the lens of SRT, our historical and social context is conceived as a set of external factors affecting our SRs, and as realities stemming from our SRs (Camargo et al., 2018).

According to Moscovici (1961), SRs are reworked when they are disseminated by the media, which act as great mediators in the production of public approval. The media instigate the construction of new SRs, or the consolidation of those already being processed in social thought (Clémence et al., 2011). SRs are produced through discourses, narratives (Moscovici, 1984), texts, cultural artifacts, and images (Jodelet, 1989; Zittoun & Gillespie, 2015) – all of which are disseminated by the media.

Several studies applying SR theory to explain the media’s influence, revealed a convergence between the press and the public’s perception over time (Bauer, 2002). As common-sense content circulates and is recycled, what is old can seem new because communication induces complex transformations in common sense (Casto & Gomes, 2005). Two of the most important systems involved in these transformations are interpersonal communication and the press (Bauer, 2002; Castro et al., 2011; Marková, 2000). SRT has several analytical tools for examining the latter’s functioning and impact. One is anchoring, a process regulated by a socio-normative meta-system and involved in classifying people, objects or ideas. The selection of anchoring categories derives from the socio-normative meta-system that organizes groups’ lives (Castro & Gomes, 2005). Another tool is objectification, or making concrete what was previously abstract (Moscovici, 1988). These two processes are not separate, but two sides of the same story (Marková, 2000), interwoven and partly simultaneous (Hakoköngäs & Sakki, 2016).

Research conducted within the social representations approach reveals how people’s opinions follow previously held positions and inter-group processes (Elcheroth et al., 2011) – a situation that has become blatantly clear through apparently apolitical issues, such as mask mandates, vaccination or climate change. This happens because our capacity to understand phenomena and their meaning relies on shared fields of knowledge that regulate where phenomena are situated within a broader social field. Political views exist on two levels: on the field of knowledge, we base them on; and on how our knowledge of specific objects reinforces, and is reinforced by our worldviews (Elcheroth, 2011).

1.2 The Present Study

This paper analyzes how traditional Brazilian media with different ideological backgrounds have contributed to constructing the discourse on political polarization in times when episodes of political violence have occupied more and more space in the papers.

2.1 Data Analysis

3.1 Descending Hierarchical Analysis

The 82 initial texts, totaling 71,782 occurrences, were deployed in 2023 text segments (TSs), 94.32% of which were retained by DHA. After lemmatization, 7725 different forms were identified (Figure 1).

Figure 1:

Political polarization in Brazil. Source: the authors.

In the first partition, the corpus was divided into two sub-corpora: A was named Emotions × Rules of Democracy (Classes 2, 1 and 3); and B was named Political Violence, meaning episodes of physical or verbal violence, according to the texts, on the part of the presidential candidate Bolsonaro (Class 4). The second partition differentiated between classes 2 and 1, which respectively indicate a relationship between an intergroup end (Class 2) and an interpersonal end (Class 1) of Class 3 – in other words, with the Internet as an important vehicle for political polarization and political violence. The dendrogram shows the distribution of the classes with details of their content.

Class 4 is opposed to all the other classes. It was associated with texts written in 2018 and 2019, by journalists or unnamed authors, and published in the Folha de S. Paulo. This class reveals how political polarization peaked in Brazil during 2018 and 2019, mainly personified by Bolsonaro, and anchoring and objectifying this violence through several episodes of violence and politically-motivated deaths in Brazil.

“(…) the capoeira teacher Moa do Katendê (…) was murdered, stabbed 12 times by a Bolsonaro supporter with whom he had an argument in a bar after the results of the first round of voting were published” (Text 26, written by a journalist, October, 2018, Folha de S. Paulo).

Class 4 also groups texts that deal with Bolsonaro’s accountability for his hate speech, underscoring the danger to Brazil when powerful people adopt such a discourse. The texts also mention the violence suffered by the candidate himself as something unacceptable in a democratic culture.

“We are aware that Brazil has a homophobic culture. When the president opens the gate, all the devils are set free (…) President Bolsonaro, who, in 2011, went so far as to say that he would prefer his son to be dead rather than gay. In October last year, after episodes of violence committed all over the country by some of his supporters), Bolsonaro released a note saying: (…) ‘I ask you to cast a null vote or vote for the opposition, for the sake of consistency; besides, the authorities will take the appropriate measures, also against the slanderers trying to harm us’, wrote the then presidential candidate, who was himself the target of one of the most violent episodes of the campaign when he suffered a knife attack in September” (Text 35, journalist, January 2019, Folha de S. Paulo).

A connection is also made between current political figures and those of the past, or even legends and myths, in an attempt to explain political violence.

“(…) Despite the fashion for describing ourselves as a peaceful people, Brazil has a violent streak. Violence is everywhere and cannot be absent from politics. It emerges in several ways. The mild ones are symbolic: Bolsonaro with Hitler’s moustache, Temer dressed as a vampire, Lula as Pixuleco. Many are more serious. Some people at a cancelled protest in Medianeira had brought dung to throw at opponents. Others near them were more successful throwing eggs, and some resorted to whipping, stoning and shooting. When bullets hit Marielle’s car and Lula’s bus, even Temer was outraged (…)”. (Text 33, professor/researcher, April 2018, Folha de S. Paulo).

Class 2 contains 339 TSs (17.77% of the corpus) and is associated with authors involved in politics. Class 2, “Democracy and Human Rights”, is related to Class 1 as it includes texts condemning political polarization and fearing serious damage to democracy. Several texts were written by women politicians from left-wing parties, and one was written by a representative of an NGO against gender violence on women in Brazil’s political world. She draws attention to the need to respect human rights in a democratic society. Political violence is presented as a confrontation between groups (men versus women; left versus right, etc.) using a defensive rhetoric that legitimizes the position of the ingroup (women) and delegitimizes the practices of the outgroup (mainly sexist, prejudiced and violent men against women in politics).

“(…) no political disagreement or protest justifies the trivialization of gender violence, a patriarchal and misogynistic practice that shatters human dignity (…) UN Women follows the Brazilian political context with great concern, and publicly calls for protection for the democratic state and rule of law. UN Women addresses Brazilian society and calls for serenity in demonstrations and non-violence in the public debates essential to a democratic management of the country’s political decisions (…)” (Text 19, journalist; March 2016, Folha de S. Paulo).

Class 1 includes texts insisting on the need to comply with political rules, to respect one’s opponents, and to adopt a set of socially-validated ethics, especially in sensitive moments like the mourning of former-president Lula’s wife. This class presents conflicts from an interpersonal angle, emphasizing that opponents are human beings, whatever the political differences. It is a call from the traditional media to abandon the conflictual stance between groups in times of crisis, and to focus on the interpersonal aspects.

“Brazilian political polarization has thus moved into dangerous territory, treading on the emotional rules relating to death (mourning and compassion). This is dangerous ground not only because it ignores one of the fundamental moral dogmas of our society, but also because, by touching on such a delicate subject, it has the potential to exacerbate moral indignation and the stereotyping of political opponents, raising people’s dehumanization and resentment to hitherto unseen levels. Denying a person compassion when a member of their family dies is a serious indication that the person is not acknowledged as human, but only envisaged as evil” (Text 43, researcher/professor, February 2017, Estado de S. Paulo).

The author of the above text could not be accused of taking sides, however, because – he also mentions how the situation was to mobilize Lula’s political supporters against his opponent:

“(…) The political conflict on emotions neither stopped at this point, nor on one side, however. Lula’s supporters exploited people’s indignation over his political opponents’ disregard for the emotional rule as a way to accuse them, sensitize people, and mobilize combativeness (…)”

Class 3 addresses the phenomenon of violence and political polarization, anchoring and objectifying this violence through the confrontation between Brazilian political groups.

“Right-wingers for the sake of convenience and staunch right-wingers are jumbled together and weaken the right. (…) The problem of the right has always been, and continues to be that it fights the left as if the left were the devil, the cause of all evils in the world. Right-wingers see the left as a monolithic block to be attacked head on, without drawing any clear distinctions” (Text 60, journalist, November 2015; Estado de S. Paulo).

The texts in Class 3 also analyze the role of the Internet as a vehicle for disseminating political polarization.

“A psychologist has written that the people who are most aggressive on the internet are helpless and frustrated: their sense of impotence and frustration makes them need to impose on other people. This is how the American psychologist Pamela Rutledge (…) judges the aggressiveness of many social media commentators in times of political polarization in Brazil” (Text 71, researcher/professor, September 2015, Folha de S. Paulo).

3.2 Content Analysis

The distribution of the articles remained relatively stable over time, peaking steeply during the run-up to the elections (second semester of 2018), with 25 articles published in 2018, and 20 in 2019. Off 82 articles, 57 were published in the Folha de S. Paulo, and 25 in the Estado de S. Paulo. The articles included 68 long texts and 14 short notes (less than 10 lines). In both papers, the texts were written mainly by experts/researchers on political sociology, politics and other social sciences (33), and journalists (31). They all strongly condemned violence and political extremism, using various arguments.

Table 1 shows that 28 articles used balanced arguments to discuss political polarization. In other words, these articles addressed the subject, criticized violence, and held both sides involved in the conflict accountable, without identifying culprits on one or other side of the political spectrum. Another 28 articles criticized political violence by pointing a finger at one side of the political spectrum, with 19 articles openly criticizing Bolsonaro for his hate speech, and as an instigator of violent acts. Another two articles criticized the politicians Dilma and Lula. The others were broadly critical of the right or the left, without citing specific politicians’ names. Finally, 26 articles addressed historical aspects of political violence, critically analyzing it in context. These articles were classified as neutral as regards the issues debated in the present; they sought explanations without pointing at any ‘culprits’ in the political arena.

Most of the publications in the Folha de S. Paulo were articles that took sides, whilst the Estado de S. Paulo concentrated more on articles taking a balanced stance on the facts. As for the authors, both newspapers gave much the same space to texts authored by researchers and by journalists. The only difference was a space made available to politicians in the Folha de S. Paulo for texts written by women on gendered political polarization and violence. Three of the four texts were written by women politicians from left-wing parties, and one was written by the president (a woman) of an NGO concerned with women.

Media productions on political polarization in Brazil are complex and illustrate Brazil’s current situation with in-depth analyses of the facts, contextualizing them historically and socially.

Our analysis shows that the two newspapers were strongly critical regarding Brazil’s political polarization, and the one-sided arguments are higher in Folha de S. Paulo than in Estado de S. Paulo. Both newspapers, and the Folha de S. Paulo in particular, criticized Bolsonaro’s hate speech. They gave more space to voices against polarization, and especially against the violence of Bolsonaro’s followers – though they also mentioned the acts of violence provoked by political agents linked to left-wing parties (mainly in texts written in 2015, when the PT was still in power). As a rule, the arguments in the two dailies were in favor of respect for differences, human rights and democracy.

4.1 Final Considerations

Of course, no analysis on articles published in the press will be able to account for all the variety of opinions and arguments on the topic of political polarization in a given community or country. It nonetheless seemed worthwhile to examine how two important voices of the Brazilian national press “gave visibility” to this phenomenon and the arguments they put forward to explain it.

A limitation of the present study lies in the small number of texts identified on the topic being analyzed. Further studies on the subject perhaps analyzing the comments of the readers in the social networks of the newspapers or with other methodologies could contribute towards the identification of potential “pathways to a change” in a country’s political polarization on both the individual and the social levels.