1. Introduction

The internet offers different discursive arenas in which people can interact, both with those who are part of their social circles as well as with unknown people, making these online spaces extensions of public spheres (Bruns & Highfield, 2016), in which contact with strangers it becomes commonplace. Social networks, “social worlds” built by different individuals with shared meanings, purposes, knowledge, understandings, and identities (Crossley, 2010), have a profound impact on people’s daily lives, changing the way they relate, build, and perceive values and how they construct meanings and senses (Recuero & Soares, 2013).

Users exploit social networks and uses it also to spread derogatory, abusive, or otherwise harmful content to people in general or specific groups (Papacharissi, 2015; Saleem et al., 2017). One of the exemplary cases of the misuse of information from users of social networks is the scandal of Cambridge Analytica, a political advisor that directed Donald Trump's digital campaign in 2016 and collected confidential information from about eighty-seven million users without their knowledge, using to disseminate degrading content, interfering in elections (The Guardian, 2019).

In the online environment, the conversation between individuals and their permanence in the reproduction of stereotypes allow the legitimation of symbolic violence to occur more easily and it can be replicated at the same speed in which it is legitimized, also allowing the dissemination of discourses full of stereotypes and prejudices (Amaral & Coimbra, 2015; Recuero & Soares, 2013). Social networks are low-cost and highly effective tools for individuals or groups to produce, legitimize and reproduce, to exhaustion, hate speech for large audiences, thus minimizing the communicative effect of these messages, which can be attitudes and conduct, and no more as speeches (Herz & Molnar, 2012; Timofeeva, 2002).

The internet facilitates an expansion of the public sphere, as it can offer a more receptive environment for those who normally do not participate in political life, instigating exploitation, creating groups, communities and making people, strangers to each other, express themselves their opinions, values, and points of view (Dahlberg, 2011; Ferreira, 2016). If we understand that the quality of democracy is related to people's engagement with politics, strong exposure to uncivil behavior generates polarization and can affect the quality of democracy, as people tend to distance themselves (Tucker et al., 2018).

In the Brazilian political scenario, there is a generalized perception of a great political polarization between left and right (Odilla, 2018), especially after the demonstrations in the streets and on digital networks in 2013. This polarization has been deepening with a series of events gaining strength, especially on social networks (Silva & Sampaio, 2018), giving “evidence that the reorganization of the right in parliament, on the streets and in the media and the offer of an extreme right candidacy in the electoral competition created favorable conditions for alignment, in the electorate, between ideology and electoral choice in 2018” (Fuks & Marques, 2020, p. 2).

The use of left and right categories to indicate political preferences dates to the French Revolution, in which delegates identified with egalitarianism and social reform sat on the left of the king and delegates identified with aristocracy and conservatism on the right. In general terms, left is related to people as the defense of social equality, heir to socialist principles, and the right linked to the defense of free capitalist markets. Still, the parties position themselves in relation to the desirable weight of state intervention in the economy, between the extreme left (full government control) and the extreme right (completely free market) (Tarouco & Madeira, 2013).

Incivility has manifold definitions and can be disrespect to the collective traditions of democracy, that is, a set of behaviors, attitudes and beliefs that threaten democracy, denying people their personal freedoms and stereotyping social groups (Papacharissi, 2004). This was visible in the 2018 Brazilian presidential election, in which there was greater polarization and aggressiveness, manifested not only in rhetoric, but also in physical aggression (Mellis, 2018). According to Maciel et al. (2018), situations of violence spread across the entire country, and no longer media and researchers can lead with it as isolated facts, but as aggressions and threats related to the polarized political scenario. In addition, reports of aggression and demonstrations of intolerance for political reasons increased on social networks, as pointed out by the Public Policy Analysis Board of the Getúlio Vargas Foundation (Angiolillo; Passo, 2018; G1, 2018).

The State of Minas Gerais, region in Brazil in which we realized this research, brings characteristic elements that can help in understanding the political identity of Minas Gerais. As pointed out by Arruda (1990), the attachment to the past is a constant in the Minas Gerais political imaginary, with the vision of today nourished by the essence of the past and the desire to accomplish in the future what has already happened. The State of Minas Gerais represents the values of duration, permanence, resistance to change, balance and conciliation with a view to avoiding ruptures, Freedom within the Order, the opposite of revolution, and tradition that resists the modern. Still, a conservative view of the world prevails, through an essentialisttraditional discourse that works as a barrier to social movements contesting the traditional order (Ramalho, 2016).

Considering the political context of the 2018 presidential elections in Brazil, particularly, by Facebook pages analyses with right-wing and far-right political bias and online incivility, this research proposes to map and analyze patterns of incivility as a phenomenon triggered in social networks online. Then, we studied fifteen Facebook pages selected that expressed support for the right and extreme right in the 2018 elections, a context marked by polarization and aggressiveness in Brazilian society. The pages were identified by searching on the Facebook social network and selecting according to the following criteria: a) public profile; b) more than a thousand participants; c) content aimed at the purpose of the research; and d) active discussions in the period from 07/22/2018, when the candidacy for the presidency of Brazil was made official, until the second round of elections, on 11/04/2018. To achieve the proposed objective, we realized digital social research, of an interpretative nature, and we analyzed the empirical material with the technique of content analysis, according to Krippendorf (2004). Then, we present approaches on incivility, the research procedures and the results achieved. Final considerations close the paper.

2. Incivility: concepts and categories

The term incivility is defined in different ways and explored in different spaces, such as in schools (KNEPP, 2012; MORRISSETTE, 2001), in the workplace (Andersson & Pearson, 1999; Cortina, 2008; Mendonça, Siqueira, & Santos, 2018), in the online environment (Coe, Kenski, & Rains, 2014; Rowe, 2014), being the object of study in fields of knowledge such as psychology, education, sociology, among others.

Incivility means disrespect for the collective traditions of democracy, a series of behaviors that threaten democracy, deny people their personal freedoms and stereotype social groups (Paparachissi, 2004; Sliter, Sliter, & Jex, 2012); acting rudely or impolitely, without consideration for others, violating the norms of respect in social interactions, including rude criticism, name-calling, disconnected speech, offensive statements, heated discussions, hate speech, harassment, aggressive comments, humiliation, outrageous claims (Andersson & Pearson, 1999; Antoci et al., 2016; Hwang et al., 2008; Jamieson, 1997; King, 2001); form of bad behavior, implying impoliteness, rudeness and disrespect towards others (Coe, Kenski, & Rains, 2014; Cortina et al., 2001), in a way that violates the norms of respect, which can lead to disconnection, breach of relationships and empathy erosion (Pearson, Andersson, & Porath, 2000; Pearson & Porath, 2005); attacks that go beyond differences of opinion, resulting in name-calling, contempt, and mockery (Brooks & Geer, 2007)

Incivility manifests itself in rude, insensitive, disrespectful, and reckless behavior with ambiguous intent to harm another individual, involving reckless words and actions that violate conventional norms of conduct (Kane & Montgomery, 1998; Pearson, Andersson, & Porath, 2000), not necessarily being a reason for any type of legal intervention (Bottoms, 2006). Mendonça, Siqueira and Santos (2018) point to a particularity of incivility, which is the volatility of roles, since a person can be a victim and, at the same time, instigator (aggressor) or victim and witness or witness and instigator (aggressor).

As for its effects, although incivility is not illegal and may not lead to catastrophic consequences or other forms of misbehavior, it has the potential to cause harm and we cannot ignore it, potentially having a harmful impact on those who suffer and those who witness the uncivil behavior (Montgomery, Kane, & Vance, 2004). Uncivil expression not only increases moral outrage but also has a detrimental effect on liberality when compared to civil expression because its detrimental effect is being most noticeable when people's views are uncivilized way attacked on the internet, diminishing the space for debate (Santana, 2014).

An important aspect in understanding incivility is the ambiguity in relation to the intention (Pearson, Andersson, & Porath, 2005; Sliter, Sliter, & Jex, 2012), since the clarity of the intention varies according to the interpretation, involving the subjectivity of the observer (Andersson, & Pearson, 1999), therefore, incivility depends on several factors, such as social position, ideology, among others (Massaro & Stryker, 2012).

The ambiguity as to whether the people’s behavior intended to be degrading, insulting, or intimidating is a key element that differentiates incivility from other types of interpersonal aggression that have a clear intention to harm the other (Desouza, 2011). The subtlety and veiled aspects that define incivility are like modern sexism and racism. In the same way that antiminority behavior is no more accepted in general societies, perpetrators of uncivil acts can easily mask their intention to harm the other person, claiming that it is just an anecdote or a joke (Cortina, 2008; Desouza, 2011).

Incivility is not related to violent behavior directly, but aggressive categories of deviant behavior that can occur, such as violence (Pearson, Andersson, & Wegner, 2001; Mendonça, Siqueira, & Santos, 2018). Violence is an intense form of deviant behavior and includes physical aggression, unlike incivility, which is less intense and excludes any physical contact (Pearson & Porath, 2004). Cortina (2008) seeks to characterize incivility to distinguish it from other acts, pointing out that it differs from other psychological aggressions when it is unclear whether the behavior caused intentionally. While incivility may have visible harmful purposes, but it can be related to other factors such as the instigator's ignorance, carelessness or personality, intent, and ambiguity.

The perspective of incivility in online social interactions emerged even before social networks (Silva & Sampaio, 2018). Since the end of the 1970s, studies have already pointed out differences between face-to-face interactions with others, especially when addressing different opinions on a given topic (Diener, 1979; Kiesler et al., 1984). Research related to incivility and the online environment had as an important landmark in the study led by Papacharissi (2004), which draws a parallel between civility and courtesy within an online democratic context. For the author, one act of incivility is as important as ten or twenty, because it attacks democratic ideals in the same way, and yet, a single act of incivility can be more severe than diverse acts together.

Incivility in the online environment is known as cyber incivility by Parker, Fritz and Jex (2015), and cybernetic incivility by Lim and Teo (2009). Both studies consider it an interpersonal stressor, which includes sending messages online in a discourteous tone, saying hurtful things, paying little attention to a request, or time-sensitive messages, such as canceling an appointment at the last minute.

Giumetti et al. (2012) offers explanations on how the internet has enhanced people's engagement in adopting uncivil behavior. The authors argue that the use of the internet facilitates the way individuals engage in uncivil behavior, whether intentionally or not, since in these environments it becomes easier to misinterpret the meaning behind the messages. Still for the authors, even if the message is not intentionally harmful, it can be interpreted as an act of incivility, since the characteristics of face-to-face dialogues, such as tone, intonation and body language are not present, which can encourage more uncivil conversations when compared to the face-to-face environment (Dutto, 1996; Hill; Hughes, 1998; Papacharissi, 2002), both because of the technologies as well as the fact that spectators cannot intervene when incivility happens (Desouza, 2011).

Facebook pages and Twitter profiles of actors and political parties, journalistic vehicles, celebrities, among others, provide a typical scenario for online incivility, since, in these settings, users can interact with strangers who signed up on the same page and, even if subscribers may have specific interests in common, they will probably be heterogeneous in terms of personal traits, preferences and modes of social interaction (Pfeffer, Zorbach, & Carley, 2014; Silva & Sampaio, 2018).

Antoci et al. (2016) characterizes online incivility as a form of offensive interaction that can range from aggressive comments, heated discussions, rude criticism, offensive allegations, hate speech and harassment. For Brooks and Geer (2007) and Coe, Kenski and Rains (2014), incivility requires going a step further and not just making a negative comment, that is, inflated and superfluous allegations that add little to the discussion. Sobieraj and Berry (2011) bring another type of incivility, the insult, which can be translate as an illegal and imprecise conduct, with the intention of diminishing a person or group. For authors, insults are uncivil, but not all incivility is an insult.

As for the operationalization of research on online incivility, Papacharissi (2004) developed an instrument to label a message as uncivil based on three questions: (1) does the discussion verbalize a threat to democracy (example: does it propose to overthrow a democratic government by force)? (2) Does the discussion stereotype (eg, associate a person with a group using labels, whether mild or more offensive)? (3) does the discussion threaten the rights of other individuals (eg, personal freedom and freedom to speak)? The author found that the most common type of incivility was the use of stereotypes as an offence.

For Massaro and Stryker (2012), the assumption that incivility dominates the online context may be incorrect, because, although the online environment facilitates rude and enthusiastic conversations, it is more marginal than dominant. For the authors, despite agreement in research on key aspects of online incivility, these are still not sufficient for empirical inferences. With that in mind, the authors point out eight categories in which uncivil discourse normally falls: (1) excessively ad hominem, which demonizes political opponents and relies on global attacks directed at character rather than ideas and conduct; (2) is false and negative about a political opponent, or is intentionally untruthful about the opponent's views, character or conduct; (3) excessively vulgar or disrespectful, or relies on excessive profanitydirected at a person (rather than an idea or institution) to gain advantage in the argument; (4) pejorative, hyperbolic, and which falsely portrays political opponents as traitors, deadbeats, Nazis, lunatics, rednecks, satanic, or unpatriotic, rather than citizens within a pluralistic political order, with whom they vigorously, or passionately, disagree on issues specific for specific reasons; (5) intentionally threatening the welfare of a political opponent, or encouraging others to cause physical harm; (6) against a political opponent with racist, sexual, religious or other epithets that a reasonable person would find extremely humiliating; (7) intentionally directed at closing "spaces of reason" and closing discourses, rather than maintaining speaking zones for further consideration on issues and policies; and (8) that intentionally denies the political right of opponents to participate equally in applicable political processes, debates or proceedings, or that denies the legitimacy of the participant, even though they have the legal right to do so.

Coe, Kenski and Rains (2014, p. 661) operationalized incivility in five ways, focusing on the notion of disrespect: name-calling (mean or derogatory words); defamation (mean words directed at an idea, plan, policy or behavior); lying (stating or suggesting that an idea, plan or policy is dishonest); vulgarity (using profanity or language that may not be considered appropriate in professional speeches); and pejorative speech (belittle the way a person communicates).

According to Gervais (2015), uncivil speech is synonym of allegations that are deliberately disrespectful and insulting, or of a hyperbolic nature. In his research, the author developed a set of categories of incivility, with the comment that violated one or more of these categories being coded as uncivil: (1) cursing, mockery and character assassination (additional superfluous adjectives and adverbs that do not add new information, but are purposefully insulting, derogatory, and condescending); (2) Misrepresentation (Use of inflammatory words or phrases that make an individual or action appear more radical, immoral, or corrupt); (3) Histrionic (Language suggesting that an individual or group is to be feared or is responsible for sadness, including intentionally exaggerated thoughts through capital letters and multiple exclamation marks). The author identified incivility based on textual communication, consolidating it into the three categories presented above: insulting language, extremist, and hyperbolic language, and histrionic and emotional.

In the context of online social networks, spaces for the production of online speeches highlight issues related to a type of incivility, which are hate speeches (Silva, & Sampaio, 2018), understood as speeches or any forms of expression that seek to promote, express or increase hatred against a person or group of people, because of a characteristic they share or the group to which they belong (Saleem et al., 2017). Ianto-Petnehazi (2012) defines online hate speech as text, audio, video, or multimedia content, usually created by non-professional or anonymous users, with the aim of intimidating or harming minority social groups (ethnic, sexual, and racial), through the resources or content hosting of the platforms. Hate speech can be explicitly hateful or veiled manifestations, hiding through statements that seem normal or rational, or the explicit exclusion of a social group or person (Rosenfeld, 2001; Moura, 2016).

al, or the explicit exclusion of a social group or person (Rosenfeld, 2001; Moura, 2016). These demonstrations produce harmful effects by violating fundamental rights, reducing the self-esteem of victims, preventing their participation in civil society activities, including public debate (Fiss, 2005; Silva & Sampaio, 2018), putting human life at risk, since that supports hate crimes and physical violence (Cohenalmagor, 2015). Online incivility carries in its theoretical foundation the bases of incivility in face-to-face relationships, changing only the way that it is arriving by messages to its victim. The crucial point we are considering in this paper based in online incivility is the characteristics and particularities that the online environment carries, such as the ease of disseminating content and interaction with a greater number of people, no matter if they are known or strange.

3. Methodological Procedures>

This research has a qualitative nature, portrayed as digital social research (Marres, 2012), since it starts from the observation intermediated by digital technologies, which allow a wide range of new practices, involving the recording, analysis, and visualization of social life (Fielding, Lee, & Blank, 2008). We opted for the social network Facebook, the most popular in the world, with more than 2.27 billion users, 130 million of which in Brazil alone. Facebook as a tool for political discussions (Stocker & Dalmaso, 2016; Saleem et al., 2017; Silva & Sampaio, 2018).

To select Facebook pages, we used the search engine, introducing keywords such as: “right”, “Jair Bolsonaro Minas Gerais”, “right Minas Gerais”, “Bolsonaro Minas”, “PSL”, “extreme right” , “conservative”, “conservatism”, “reactionary”, “ultra-right”, which are associated, in the Brazilian political context, with the Social Liberal Party (PSL) candidate who defends right-wing and extreme-right policies. Such political alignments are as typical of a capitalist economic model with appreciation and traditional moral precepts (Codato, Bolognesi, & Roeder, 2015) and by an association with conservatism, which raises a flag to combat corruption, a repressive ideology, the cult of violence, intolerance, and the appeal to the military (Löwy, 2015). With this understanding, only the pages and groups with the following criteria were surveyed and pre-selected: a) be public; b) more than a thousand participants; c) content aimed at the purpose of the research; and d) active discussions in the period from 07/22/2018, when the candidacy for the presidency of Brazil was made official, until the second round of elections, on 11/04/2018. After this initial analysis, we selected data of fifteen pages to this research, which are in Table 1.

Table 1 – List of pages selected for the survey

Source: Elaborated by the authors

We realized data scraping by Netvizz, obeying the limitations of the social network and the application for this type of collection, such as the anonymity of users and the full availability of comments, given the commercial nature of the platform and Facebook's privacy policy (Silva & Sampaio, 2018). This stage of the research generated a total of 1,909 posts and 102,116 comments. Considering the purpose of the research, the filter with terms that are associated with incivility resulted in 3,019 comments. To reach the research corpus, we selected the comments that met the established criteria, discarding those that contained information about people, cities, advertising content, emojis, civil content, hashtags, and other types of content that did not fit.

The analysis of the material also was complementary by the free software IRAMUTEQ. We did a second manual review of the comments and discarded ninety-seven of them as they only contained encouraging emojis like clapping. After the second scraping, we submitted the 2,922 comments to the Iramuteq software, which is a significant resource for qualitative research in the handling, systematization, organization, and presentation of data.

To apprehend the specificities and discuss the incidence of incivilities expressed in the online environment, in addition to the textual analysis in the Iramuteq software, the comments selected were also analyzed by our research team from the perspective of content analysis (Krippendorff, 2004), which allows identifying, analyzing, and reporting standards, organizing, and describing the research corpus. The content analysis was operationalized based on the framework of Krippendorf (2004), which considers the following conceptual components: a) a textual corpus - which in this research are the selected comments; b) a research question to be answered with the textual corpus – which patterns of incivility can be found; c) a context to give meaning to the corpus, since the texts acquire meaning depending on the context – 2018 presidential elections in the Brazilian context; d) an analytical construct that operationalizes – the categories of incivility proposed by Massaro and Stryker (2012); e) inferences that are intended to answer the research question – the abductive logic for interpretation as presented in the results section; and f) validation evidence – for Krippendorff (1980, 2004), after describing the criteria for a good measure of reliability for coding data, Krippendorff’s alpha can be as the standard reliability measure.

4 Patterns of Incivility on Facebook

The results presented here recognize the growth of uncivil discourse in the online environment, not only in quantity, but also in its dynamics and ability to reach its targets. The 2,922 comments that make up the data of this research needed to be categorized by us based on the literature review (Massaro & Stryker, 2012).

Incivility 1 - Speech that is excessively vulgar or directed at a person or group

The first category, which contains comments with vulgar or disrespectful content directed at a person or group, was numerically the most representative, corresponding to 2,263 comments and 77.45% of the total. These comments included insults, stereotypes, sarcasm, irony, name-calling, vulgarity, LGTBphobia, misogyny, stigma, prejudice, lies and any other type of incivility, intolerance or hate speech directed at someone or a group.

Within this category, some themes are recurrent. One of these themes would be related to ideas expressed in an uncivil way about human rights, the comments bring false ideas about what human rights would be and ironies (called “right of the brothers”), within a pejorative context.

In this category, comments such as “human rights, the dung of vagrancy”, “it encourages violence is a father to see his son dying in hospital lines and nobody does anything, bunch of bums, these human rights are rubbish, “you guys are rubbish they only defend the rights of brothers, the family has no value for you” and “human rights rubbish will end your magnate”.F

Comments expressing incivility through discrimination based on sexual orientation or gender, using insults or irony, are also common. These situations can be illustrated in statements such as “she will not kill gay; he hunts wild boars” and “first, the fagot said that he even takes a gun, if necessary, of course he takes a pistol all day lol, and the naughty woman quoting the bible, the scoundrel had to wash her mouth to talk about the scriptures. bunch of pigs”.

Attacks on feminists are also common in the comments, in which the feminist movement is disqualified through false and generalist ideas, in addition to insults and name-calling, such as “little feminist whores”, “poor girl was molested and became a feminist mkkk”, “better than a feminist hairy sure is” and “of course! starting with the armpits!!! want to make a feminist run? just show her a quick beard!!”.

Cursing, insults or offenses, incivilities expressed gratuitously towards an individual or group or for not agreeing with a position different from that expressed by the right and extreme right: “garbage! Devils! scoundrels! bad professionals! Incompetents!”, “you should worry about learning how to write illiterately”, “imbecile muggle sucker”, “idiots”, “sniffing, drunk and sucker of the Rouanet law (Brazilian law named after Sérgio Paulo Rouanet whose role is providing monetary funds for use in art and culture)!”, “these people are trash, all bums” and “scumbags, disgraces, damn, rubbish.”

Incivility 2: Pejorative and hyperbolic speech against political opponents

In the second category, comments with pejorative and hyperbolic content directed at political opponents are grouped, corresponding to 351 comments and 12.01% of the total (second place in numerical representation, well below the 77.45% of the first category). The comments are curses, insults, vulgarities, sarcasm and irony, forms of prejudice, intolerance or hate speech directed at opponents (parties or politicians) who are not considered by the participants of the discussion as supporters of right and extreme right speeches.

For the most part, the comments expressed here directly attack people and parties linked to the left, but not as a form of opposition or discussion of political or ideological ideas, but personal attacks. As, for example, those directed at Dilma Rousseff, who in this context was running for the position of senator for the State of Minas Gerais: “please do not vote for the president...”, “they say that the president has a doctorate. where did she buy her diploma?”, "and to think that this disgrace was once president of this country, and there are many miners who will vote for this tapir called dilma Odebrecht”, “our future senator tapir: o minas Gerais. .. because?" and “you guys are really crazy, you made me proud to get that slut out of dilma”.

Comments with incivilities directed at Luiz Inácio Lula da Silva, and likewise, not as opposition to ideas, but rather, just personal, and gratuitous attacks. This can be seen, for example, in "Lulalau 157 Condemned Vagabond", "The world is happy to see Trash Squid arrested!" and “crying is.. just charming donkeys and brainless lullabies, no kkk”. There were also uncivil comments directed at presidential candidate Fernando Haddad. As in the previous situations, they were directed at the person, and not at ideas or government proposals: “this haddad who is president kkkkkk is still this crook”, “and haddad and you better? resign your candidacy because you? Go to jail, you worm Sérgio Moro, you'll get it” and “You are a dangerous thief Haddad”.

Comments directed at the Workers' Party and its sympathizers, and like the others, are made up of out-of-context and generic insults: "outside the PT, they're going to teach you to pick up your dick, your dog's left", "just left rubbish, used public offices to campaign for the corrupt, they are rubbish that add nothing to society, they only think about themselves” and “Bolsonaro is having lunch with the elite people in a 5-star place, that’s how the left-wing PT people are going to say”.

Incivility 3: Intentionally threatening and/or violent speech

The third category, which covers comments with threatening content and inciting violence, corresponded to 181 comments and 6.19% of the total. The comments contemplated in this category carry a threatening tone towards people and groups or that, in some way, incite physical violence. Most of the comments are about killing and putting an end to a certain person or group, being especially related to issues of carrying weapons (used as a campaign proposal by candidate Jair Messias Bolsonaro) and some statements by the candidate for governor of Rio de Janeiro Wilson Witzel.

eiro Wilson Witzel. Situations that referred to the military regime, alluding without historical foundation in “what a good time it was” and asking for its return are present in the comments. Such situations carry a tone of threat to democracy and are evidenced in examples such as “military civic intervention now” and “I'm sorry but I'm from the military regime. it was the best time for someone who wasn't a criminal. but the military did little. they should have killed them all.” Other comments are threatening and violent, being directed at politicians linked to the Workers' Party (Dilma Roussef and Luiz Inácio Lula da Silva), as in “leave free Lula dilma josé dirceu and all his gang. they should have killed them too for us to be a free country. and only those who are criminals are afraid of the military.” and "I think it's better to kill these two rubbish dilma and lula.".

Incivility 4: False and negative speech about political opponent

The fourth category, referring to false and negative content about political opponents, had 127 comments and represented 4.35% of the total. This category concentrates on comments that carry false or negative information about people or parties considered to be opponents, such as, for example, the distribution in schools of the gay kit by then-candidate Fernando Haddad and the Workers’ Party: “I started to admire him [Bolsonaro] ] when he fought the gay kit that taught pornography to children”, “show me a video like that of the snack thief alckmin or father of the gay kit bad, or cel”, “so the fagot there author of the gay kit will naked protect only if it is themselves" and "haddad only implemented the gay kit, a deviant, does not respect Brazilian families."

Comments that associate Brazilian artists with the use of public money to campaign in favor of the candidate Fernando Haddad or the Workers’ Party, as in “you have to do this to show the face of these half-a-bowl artists [sic] bought with public money to campaign dirty for communism.” and “a piece of junk that benefits from the Rouanet law...”.

Within this category, there were also situations involving the vice-candidate Manuela D'Ávila, with false and negative ideas, as in “I changed my vote because Haddad's vice-candidate wants to spread the march of LGBT sluts across Brazil and the booklet for children. Now I am Bolsonaro for order and progress!” and “this manuela is a walking joke lol”.

Comments propagating false ideas, such as “gender ideology”, perceived in “against this damn gender ideology, which wants to end the traditional Brazilian family”, “if we speak the truth, a thousand want to hunt us down, for being with Bolsonaro and because many want to stone us! but one thing I say is against human rights, I am against gender ideology!” and “en wanting to shove gender ideology down my throat and etc.” were also common in the analyzed material.

5. Discussion: Incivility in the Comments and Uncivil Strangers Online

The analysis of these patterns pointed out that the social network Facebook constitutes a space for transmitting political opinions, disseminating ideas with high levels of incivility, whose attitudes and behaviors challenge democratic deliberation (Gervais, 2015). The lack of social ties in social networks encourages verbal attacks and uncivil behavior, causing people not to contemplate the values, beliefs, and preferences of others (Hill & Hughes, 1998; Papacharissi, 2004; Rowe, 2014; Maia & Rezende, 2016). And normally, this incivility is directed towards individuals who are not directly involved in the discussion (Rowe, 2014), not bringing contributions.

The research results lead us to highlight three considerations that point to its potential theoretical contributions. The first one refers to the figure of the “online uncivil stranger”, which was not part of the problematization phase of the research, however, emerged from the analysis of the material. In the public sphere, where space must be shared, it is common for strangers to meet whose objective information is not known, which makes the quality of interactions difficult. However, when Sennett (1999) discusses the decline of the public man, he speaks of the withdrawal of public culture due to the development of a personalistic individuality, which has become a priority in social relations. For the author, the reclusion of the private space and silence when meeting strangers are strategies used in this withdrawal, a way of remaining detached from society. However, this is not the case in the online environment.

Online social networks, here in particular, Facebook, prove to be propitious spaces for diverse manifestations, offering versions of events, projects, and worldviews, in the online interaction of known and strange people. As exposed by Maia and Rezende (2016), when they are exposed to conflicting values and opinions, Facebook members, instead of listening carefully to opponents and engaging in argumentative exchange to defend their positions, seek to expel the “intruders” and expressed offenses without embarrassment.

The second consideration concerns the consequences of the propagation of incivilities in the online environment. Comments characterized as hate speech, offenses, curses, vulgarities, or violence bring social consequences, not only for those involved, as they contain the challenge of human rights and the threat to democracy and individual rights (Papacharissi, 2004). The social network in question cannot control the propagated speeches, despite having policies and guidelines against any type of hate speech or intolerance, this implies the naturalization of violence, making it more systemic and allowing its legitimation and replication (Recuero & Soares, 2013).

The third consideration refers to the political polarization in the Brazilian scenario, which, by highlighting the differences between the political alternatives, encourages “a more ideological response from the voter” (Fuks; Marques, 2020, p. 418). The approach we adopt only assumes that contexts that highlight the differences between political alternatives stimulate a more ideological response from voters. Comments demonize political opponents, directing them at personal attacks and not at government ideas, conduct, and programs, but with distortions, lies, insults and defamations (Coe, Kenski, & Rains, 2014). The political-partisan opposition served as a subterfuge for uncivil speeches to be more tolerable and reproduced against any group or individual that was against the ideals of people with speech aligned to the right and extreme right.

As pointed out by Silva and Sampaio (2017), incivilities reflect the polarization that has been established in the country and the damage to democracy resulting from the lack of tolerance and mutual respect, safeguarding itself in aggressive, sarcastic, and ironic arguments. Political polarization increases when incivility is present in online comments, with less liberal attitudes that instigate hostile emotions towards those who think differently about a certain subject (Hwang et al., 2008; Borah, 2014). Uncivil comments may have been more tolerated and reproduced on these right-wing and extreme right-wing pages because they are users with similar political ideologies, that is, the homophily between these users allows for greater engagement, encouragement, and tolerance of incivility.

6. Conclusion

In this research, the objective was to map and analyze patterns of incivility as a phenomenon triggered on online social networks, on Facebook pages that expressed support for the right and extreme right in the 2018 elections, a context marked by polarization and aggressiveness in Brazilian society. The content of Facebook pages in Minas Gerais linked to the right and the extreme right brings discussions that verbalize a threat to democracy, attribute stereotypes or threaten the rights of individuals, thus being able to be considered as containing uncivil discourse, reinforcing the forms of evaluation idealized in the literature.

The political context considered serves as a background for incivility to be more tolerable and reproduced against any group or individual that was against the ideals of people with speech aligned to the right and extreme right, even if it is a stranger. When describing the research results, we identified the figure of the “uncivil stranger in the online environment”, which was configured in a relevant research finding, since, as online social networks are spaces of interactions, and, particularly, the social network Facebook it is a “network of friends”, forms of sociability should consider the rules of proper behavior when meeting strangers.

The research points to the growth of uncivil discourse in the online environment, not only in quantity, but also in its dynamics and ability to reach its targets. Although the comments are not as expressive numerically, a relevant aspect is the engagement of the users of these pages in these comments, that is, even though there is no comment directly expressing an act of incivility, sharing, or demonstrating positive reactions to these speeches It's a way to encourage that kind of behavior.

The research contributes to the understanding of how social networks, in particular Facebook, can encourage the existence of incivility discourses, thus bringing empirical knowledge about online incivility in the Brazilian political context of the right and extreme right. Another contribution is to increase the visibility of the subject, allowing reflection on uncivil practices, their specificities and their possible consequences in the organizational and social context, which also makes use of the virtual space. In this sense, surveys that consider uncivil manifestations of other party-political ideologies could bring other categories, in addition to being useful for comparing patterns of online incivility.