Political narratives in dispute on Twitter: the (de) construction of Lula’s public image in the context of the Brazilian political crisis of 2016

Abstract

This article investigates how the public image of Lula, Brazil’s former president, was (de)constructed during the Brazilian political crisis of 2016 established as a result of Dilma Rousseff’s impeachment process. In order to do so, we analyzed the circulation of Twitter narratives that mentioned the keyword “Lula” and that were published during four contexts that occurred in that year, namely: the social demonstrations in favor of the impeachment that occurred on March 13; the social protests against the impediment that took place on March 18, the impeachment vote in the Chamber of Deputies, on April 17, and in the Federal Senate, on May 12. Among the main results, we highlight the narrative clash observed from an intense dispute over the meaning of the term “coup” that was used both in deconstructive and constructive reports related to the public image of Lula.

Keywords
Lula; Public image; Impeachment; Political crisis; Twitter

Resumo

O presente artigo investiga como se deu o processo de (des)construção da imagem pública de Lula, ex-presidente do Brasil, durante a crise política brasileira de 2016, ocorrida em decorrência do processo de impeachment de Dilma Rousseff (PT). Para tanto, foi analisada a circulação de narrativas no Twitter que mencionavam a palavra-chave “Lula” e que foram publicadas em quatro contextos ocorridos naquele ano, quais sejam: as manifestações sociais a favor do impeachment que se deram em 13 de março; os protestos sociais contra o impedimento ocorridas em 18 de março, as votações do processo de impeachment na Câmara dos Deputados, em 17 de abril, e no Senado Federal, em 12 de maio. Entre os principais resultados, ressaltamos o embate narrativo observado a partir de intensa disputa pelo sentido do termo “golpe”, que foi utilizado tanto em relatos desconstrutivos quanto construtivos relacionados à imagem pública de Lula.

Palavras chave
Lula; Imagem pública; Impeachment; Crise política; Twitter

Resumen

El presente artículo investiga cómo se dio el proceso de (des)construcción de la imagen pública de Lula, ex presidente de Brasil, durante la crisis política brasileña de 2016 establecida a raíz del proceso de impeachment de Rousseff. Para ello se analizó la circulación de narrativas en Twitter que mencionaban la palabra clave “Lula” y que fueron publicadas durante cuatro contextos ocurridos en aquel año, cuáles son: las manifestaciones sociales a favor del impeachment que se dieron el 13 de marzo; las protestas sociales contra el impedimento ocurridas el 18 de marzo, la votación del proceso de impeachment en la Cámara de Diputados, el 17 de abril, y en el Senado Federal, el 12 de mayo. Entre los principales resultados resaltamos el embate narrativo observado a partir de intensa disputa por el sentido del término “golpe” que fue utilizado tanto en relatos desconstructivos como constructivos relacionados a la imagen pública de Lula.

Palabras-clave
Lula; Imagen pública; Impeachment; Crisis política; Twitter

Introduction¹

The image of Luiz Inácio Lula da Silva – also known as “Lula” – has been present in the Brazilian collective imagination for nearly four decades. He not only led some of the biggest labor strikes ever reported in the country (from 1978 to 1980) but he was also one of the main founders of the Workers’ Party (Partido dos Trabalhadores – PT) in 1980. At that time, Lula stood out as the Brazilian working-class hero as well as a fundamental political persona in the (re)democratization agenda of the country, during protests that became popular as “Diretas Já”, which took place from 1983 to 1984.

Lula ran for president for three elections in a row (1989 – first direct elections in Brazil since its redemocratization –, 1994 and 1998) and he was finally elected in 2002, which allowed him to set his public image as the core of the political scenario in the country. It’s important to emphasize that, in 2002, the PT leader reportedly became the president with the highest number of nominal votes registered in a Brazilian direct election so far² - such fact has endorsed his prominence in the Brazilian public imagination.

Due to several corruption scandals that became popular as “mensalão” – reported in English as a big monthly allowance or monthly stipend – which were in the spotlight in 2004 (2^nd year of Lula’s first term), the image that the voters had of the party was harmed. Taking such damage into account, one of the main communication strategies for Lula’s reelection campaign (2006 elections) was to establish a temporary distance between the president and the party. Due to this withdrawal, Lula’s political personalism was intensified, which highly contributed to the rise of Lulismo (Lulism)³.

Despite massive coverage of the corruption scandals and the hard criticism addressed to Lula from the country’s major media outlets, he was reelected in 2006. By that time, Lula obtained a total of 58 million votes – nearly 6 million more in comparison to 2002 and a new election record in the country was established⁴. In 2010, Lula left the Brazilian Presidential chair with a historical approval rate of 80% regarding his presidential period and 87% personal approval rate, according to a survey by Instituto Brasileiro de Opinião Pública e Estatística (Ibope), conducted by Confederação Nacional da Indústria (CNI) in December 2010 (CONFEDERAÇÃO NACIONAL DA INDÚSTRIA, 2010).

All the popularity acquired during his two terms was used to support Dilma Rousseff⁵, the PT candidate who won the Brazilian Presidential elections in 2010⁶, to whom Lula transferred his political prestige. The public figure of Lula was intensified again in 2013, during the 3^rd year of Dilma Rousseff’s term, when social protests took over the Brazilian streets nationwide.

While analyzing Lula’s visibility through the context of these protests, which boomed in June 2013, a process of deconstruction of Lula’s image is evident. Therefore, we highlight a study done by Recuero (2014) regarding the contents posted by media outlets on Twitter – a microblog which enables posts with a maximum of 280 characters – from June 3^rd to June 20^th, 2013. This investigation demonstrated that mainly in the official accounts of major media outlets such as Folha de S. Paulo, Uol news and R7 news, Lula’s name was associated with the protests. It is also relevant to highlight the academic work showcased by Malini et al (2014), who analyzed hashtags⁷ posted on Twitter accounts from June 15^th to the 17^th, 2013. Based on the data collected during the research, Malini el al (2014) stated that #lulaéocancerdobrasil (Lula is the Brazilian cancer) was one of the main hashtags used by social media users.

In March 2014, Lula’s public image also took center stage in the Brazilian political context with the beginning of the Operation Car Wash – in Portuguese, known as Operação Lava Jato. The Operation Car Wash investigates a wide corruption scheme originated at Petrobras (Public Oil & Gas Company). Once again, the images of both Lula and PT were damaged with negative evaluations.

Even with this reputation crisis, which started during the 2013 protests and was intensified by the Car Wash scandal, PT was successful with a fourth victory in a row as Dilma Rousseff was reelected in 2014⁸. After her second inauguration, in 2015, the PT government not only had to face a reputation crisis but also an unstable economic situation, which led to a new set of protests nationwide.

The major propelling force behind these protests was the criticism addressed to PT, especially the demand for the president’s impeachment. In addition to that, it’s important to emphasize once again the importance of Lula’s figure in the country’s political context. In regards to this scenario, sociologist Scartezini (2016) shows that the 2015 demonstrations presented a strong anti-Lulism appeal, as the author also stated that the chant “Lula never again” was widely shouted by the majority of protest-goers.

It’s also relevant to highlight one big symbol of such protests: a giant inflatable toy that portrayed Lula as an inmate, a reference to a possible criminal conviction. This toy was nick-named “pixuleko” – this name has its origin in one of the Operation Car Wash investigation stages.

Alongside this context of unhappiness, on October 7^th, 2015, the country’s Court of Auditors (in Portuguese, Tribunal de Contas da União - TCU) decided unanimously to reject the accountancy presented by Dilma Rousseff’s team at the end of her first term, alleging fiscal responsibility crime. As it has been explained by Glasenapp and Farias (2016), TCU’s decision became the main argument for the impeachment request against Dilma Rousseff, which was elaborated by lawyers Hélio Bicudo, Miguel Reale Júnior and Janaina Paschoal, who had full support from opposition parties, such as PSDB. This request was submitted to the Deputy Chamber President, Eduardo Cunha (PMDB), on October 21^th, 2015 and then analyzed by several congressmen who decided to carry on with the accusation. Eduardo Cunha authorized the lawsuit to begin on December 2^nd of the same year.

The establishment of the president’s impeachment lawsuit caused several opinion divergences regarding the legality of such lawsuit. This controversy generated impeachment-support protests on March 13^th, 2016 – which became popular as “manifestações pró-impeachment”. On the other hand, protests against the impeachment occurred on March 18^th of the same year – this one was named as “manifestações contra o golpe” (protests against the coup d’état). In both contexts, the public image of Lula was widely mentioned by protest-goers: he was either accused of being responsible for the political crisis that the country was going through or portrayed (Lula as well as Dilma) as victims of a coup.

In regards to the protests that were held throughout 2016, it is important to mention, once again, the intense media exposure acquired by Lula in specific moments, such as: a) the coercive conduction of the former president, on March 4^th, 2016, determined by the Prosecutor’s Office of São Paulo, during the 24th stage of the Operation Car wash investigation; b) the operation in which the Federal Police of São Paulo state indexed 136 items received by Lula during his two terms as a president – this task happened on the same date, on March 4^th, 2016, as it was also part of the 24th stage of the Operation Car wash investigation ; c) Lula’s pre-trial prison request submitted by three prosecutors from the Prosecutor’s Office of São Paulo – Cássio Conserino, José Carlos Blat and Fernando Henrique Araújo –, on March 9^th, 2016; d) the nomination/appointment of Lula as the Civil Cabinet minister on March 17^th, 2016, during Dilma Rousseff’s term, after Jaques Wagner left the position (such nomination was denied by the Supreme Court Minister, Gilmar Mendes); e) Lula’s speech during major protests against the impeachment lawsuit on Paulista Avenue, São Paulo city, on March 18^th.

According to a survey by Instituto Datafolha, published on April 11^th, 2016, regarding the intention of vote for the 2018 presidential election, Lula was mentioned by respondents as their favorite candidate to win the election. Four scenarios were presented: in three of them, Lula shared the leadership with Marina Silva (Rede Sustentabilidade party) – in one other scenario, he was pointed out as the main candidate⁹. The same study measured the electoral rejection rate of all names that were being surveyed. Despite Lula’s popularity, his name obtained the highest rejection rate (DATAFOLHA, 2016). Therefore, it’s possible to say that there was an uncertainty regarding Lula’s figure amongst Brazilian voters.

Based on this information, it is our assignment to enquire: who have been the most influential characters either in building or deconstructing the former president’s public image during such period? What were the main resources that were used as strategies by both groups?

The main data sources for empirical analysis are the narratives that were posted on Twitter throughout the four contexts previously mentioned (using “Lula” as a key-word). It’s important to emphasize that the narrative concept used in this article is based on the theory proposed by the Lithuanian linguist and semiologist Greimas (1973) who, after being influenced by the Structuralist approach, aimed to prove the existence of a narrative structure – also known as narrativity – which may be established and found in any kind of text.

It’s also important to highlight that Twitter was chosen as the research environment because this article’s aim is to understand the circulation and the visibility of such narratives and public images in contemporary age, so it’s also imperative to understand the transformations occurred due to the heavy and constant use of internet worldwide, as well as the importance of mass communications in spreading political narratives.Internet and social media channels are major tools that enable the creation of connective networks and public spaces that can be used either for visibility or conversations about public subjects. As it is known, in such “places”, audiences can consume and share information quickly, as well as discuss and debate diverse themes in order to possibly obtain concrete feedback on political actions due to the extensive and increasing exposure that can be generated online. All this architecture of public space turns out to be new forms of perceiving politics, especially when it comes to day-to-day conversations regarding political matters.

For instance, Twitter has become an important research tool worldwide regarding the sharing of real-time information and also due to its ability to politically engage users considering the vast amount of social manifestations that occurred in the 2000s – such as the protests in Iran and Northern African and the Middle-eastern revolutions, in 2010 and in 2011, known amongst the western journalists as “Arab Spring”.

As it creates the means to enable the circulation of narratives and counter-narratives, which are created by their own users, as well as making these narratives reach other audiences, the microblog stands out as an extremely relevant social media tool nowadays. Santaella e Lemos (2010) state that the microblog enables the intertwining of information flows and a real-time collaborative design, as it changes and accelerates global moves of a collective mindset (SANTAELLA; LEMOS, 2010, p. 66). In addition to that, they categorize Twitter as a social and political “thermometer”, as a civil resistance tool, a cultural stage and an arena for continuous conversations. (SANTAELLA; LEMOS, 2010, p. 66).

Methodological procedures

Firstly, it is important to mention that the data was obtained with the use of a tool called Netlytic¹⁰, which was configured to collect messages (or tweets), previously published on Twitter in March, April and May of 2016 – which contained the keyword “Lula”. The data collection period chosen for this research refers to 4 meaningful contexts/moments considering the Brazilian political crisis of 2016: the social protests in favor of Dilma Rousseff’s impeachment, on March 13^th; the social protests against the coup d’état, which took place on March 18^th; the voting session to support the establishment of the impeachment lawsuit, which was held on April 17^th, at the Deputy Chamber/House of Representatives; and, then, the voting session on May 12^th at the Brazilian Senate, which resulted in the suspension of the president for 180 days¹¹.

Considering such time range (March-May), it’s noticeable that an excessive quantity of information was collected – nearly 1,705,117 tweets were gathered each month, in other words, a total average of more than 5 million tweets. The remarkable number of posts would hinder an efficient validation of this research’s results. Therefore, the method that was adopted in this article is the intentional sampling. According to Fragoso, Recuero e Amaral (2013), the aim is to understand and select the most meaningful elements from a certain range of posts; this sampling method is based on the researcher’s intention to define goals so they are able to answer their research questions.

That said, as of the pre-analysis of such data, we chose to look into specific dates. The tweets used for analysis were the ones which mentioned the keyword “Lula”, posted on March 12^th and 13th; March 17^th and 18th; April 17^th and 18th; May 12^th and 13^th – these dates refer to 2016¹². Such approach reduced the total corpus of analysis to 437,286 tweets. The empirical analysis of this data was split into two stages: 1) the first one consisted of the social media analytics (SMA), based on the idea presented by Recuero, Zago and Bastos (2015), which enabled this current piece to investigate which profiles of this microblog showcased themselves as more influential in the circulation of the narratives about Lula – again, considering the analyzed contexts. The influential rate was established due to the quantity of connections received by each profile (a higher rate of engagement/access), during the circulation of these messages in each one of these contexts. It’s important to emphasize that these profiles (the ones with a wider engagement/access) performed better in terms of quotation/tagging and sharing (retweets); therefore, they are considered influential profiles (or central ones) on the internet. Based on that conclusion, these profiles were classified into six categories¹³: a) political leadership; b) media outlets; c) news blogs; d) opinion leaders; e) activism; f) regular users.

Throughout the SMA method, it was possible to comprehend the way these profiles gather around the conversations in the analyzed contexts, also considering the modularity which enables the cluster identification – this term refers to users who are deeply intertwined and weakly connected to the rest of the network (RECUERO; ZAGO; BASTOS, 2015, p. 84). The clusters, also known as groups, are mainly formed, as members have the same interests in specific subjects, which supports the understanding of how information circulates on Twitter. This article investigated five main clusters which were identified in each of the four analyzed contexts, powered by a software called Gephi¹⁴. All the data collected by Netlytic was exported to Gephi so it could convert it into graphics representing these networks (Figures 1, 2, 3 and 4).

Figura 1
Full Network (Context 1)

Figura 2
Full Network (Context 2)

Figura 3
Full Network (Context 3)

Figura 4
Full Network (Context 4)

The second methodological phase refers to the analysis of narratives, which was based on the content analysis method proposed by Bardin (1977) and Bauer (2008). We concluded that the analysis of tweets would obtain better results from the representativity rule (BARDIN, 1977). Therefore, we verified which terms or syntactic units (BAUER, 2008) were the most frequently mentioned in the corpus of analysis. The four empirical contexts were examined separately, as it was done in ARS considering the most repeated terms in each scenario.

This selection of the most mentioned terms was made based on what Bardin (1977) calls computer processing or ordering, due to recurrence. In order to achieve this, we relied on a computer program, AntConc¹⁵, able to quantitatively check the syntactic units that were most often mentioned in the content of these tweets. With this analysis, it was possible to observe the main themes used by Twitter users when seeking to deconstruct or build Lula’s public image.

The four following tables (Tables 1, 2, 3 and 4) show the results obtained in regards to the most mentioned terms in the tweets published in each of the four analysis contexts¹⁶.

Table 1
Context 1’s most frequent terms: social manifestations in favor of the impeachment

Term	Frequency		Term	Frequency
Dilma	12837		PF	2368
Brazil	8146		arrested	1893
prison	6370		act	1884
PT	4642		PM	1826
former president	3999		Car Wash	1786
vault	3975		@VEJA	1749
Globo	3177		MP	1706
request	2760		Moro	1687
support	2669		prosecutors	1580
SP	2651		defence	1532

• Source: developed by the authors.

Table 2
Context 2’s most frequent terms: social manifestations against the coup d’état

Term	Frequency		Term	Frequency
Dilma	23324		judge	3192
minister	14431		Gilmar	2937
Brazil	10846		phone tap	2899
inauguration	9008		PT	2930
Moro	6241		coup	2572
Civil Cabinet	5038		government	2331
president	4187		justice	2214
STF	4023		injunction	1950
nomination	3718		Paulista	1941
suspension	3551		conversation	1849

• Source: developed by the authors.

Table 3
Context 3’s most frequent terms: voting session in the House of Representatives

Term	Frequency		Term	Frequency
Dilma	55910		Vote	3141
Brazil	9786		deputies	3081
impeachment	9767		STF	2984
PT	8793		government	2785
Cunha	5623		Venezuela	2782
defeat	4271		@blogdojefferson	2263
president	4018		jail	2256
Temer	3415		humiliating	2021
minister	3396		#impeachmentday	1720
coup	3216		arrested	1680

• Source: developed by the authors.

Table 4
Context 4’s most frequent terms: voting session in the Senate

Term	Frequency		Term	Frequency
Temer	12482		Coup	3062
Ministers	9739		STF	2761
Car Wash	9223		impeachment	2125
Brazil	8086		@genpenaloza	1772
PT	6932		People	1710
Investigated	6650		Jail	1690
President	5133		Inauguration	1642
Moro	4466		@dilmabr	1637
Legislative immunity	3778		arrested	1634

• Source: developed by the authors.

We also point out that the study of the material (the collected tweets) was made intentionally. Thus, we chose to select certain examples of publications that allowed us to perform a deeper analysis of the narrative content published in each of the observed themes.

The political narratives in contention on Twitter

Considering the social media analysis that was established on Twitter during these four contexts of the Brazilian political crisis, from March to May 2016, which were specified on the Methodology section of this article, we could observe the high degree of influence and, therefore, the centrality of the hegemonic media in the circulation of narratives about Lula in the analyzed contexts.

The relevance of the media outlets occurred due to the fact that the narratives with a higher sharing rate – that is, the ones that obtained referability, visibility and credibility (RECUERO; ZAGO, 2012) – were published by media profiles which belong to major news conglomerates in Brazil, such as Veja magazine (@VEJA), O Estado de S. Paulo newspaper (@estadao), Folha de S. Paulo newspaper (@folha), the G1 online portal (@g1), and also by Journalistic Blogs’ profiles associated with these companies, such as Ricardo Noblat - O Globo (@blogdonoblat) – and Felipe Moura Brasil - Veja (@blogdopim). The journalistic blog created by Diogo Mainardi and Mário Sabino, called O Antagonista (@o_antagonista), was also observed as one of the most visible profiles in the circulation of information about Lula, considering the analyzed period. We also emphasize that, among all the main profiles that belong to opposition groups to Lula, two opinion leaders’ profiles were noted: one belonging to the comedian Danilo Gentili (@danilogentili) and another one belonging to the singer and songwriter Lobão (@lobaoeletrico).

The decisive influence of these main/hegemonic outlets in the process of circulating information about the former president in the microblog reinforces the position addressed by Recuero (2011, p. 15). According to her, even with much more direct participation of the audience, the actors in social networks on the Internet still rely on outlets and media institutions to legitimize, give credibility, organize and filter information. Thus, despite the fact that these actors – as well as common users – may contribute to spread information on the internet by, for instance, retweeting content, the media outlets still generate visibility and relevance to some narratives.

A second point to which we would like to draw attention is regarding the social media analysis and it refers to the fact that the circulation of such reports on Lula in the analyzed contexts is based on the polarization between groups which posted narratives to construct the public image of the former president (supporters) and the groups which aimed to deconstruct his public image (opponents). However, this polarization didn’t happen equanimously, as there was a quantitative predominance of groups circulating narratives of Lula’s public image deconstruction.

Only in the fourth context of analysis – during the period in which the impeachment voting session was held at the Senate – the predominance of the former president’s supportive groups was observed. It is important to mention that the predominance of the narratives to construct the political public image of Lula, which happened only in the fourth context, befits the fact that the outlets with a greater amount of mentions in this context, contrary to what was observed in the first three contexts, didn’t belong to media outlets, but to PT-linked leaders, such as Dilma Rousseff’s page (@dilmabr) and the Instituto Lula’s one (@inst_lula) – the latter was widely mentioned by Twitter users as if it were Lula’s official profile (@lulapelobrasil), which indicates the lack of knowledge from the audience regarding the official communications channels of the former president on the microblog. It’s also necessary to inform that the centrality of these two profiles - (@dilmabr e @inst_lula) -, in the studied context, does not necessarily indicate a greater visibility of the narratives to construct the public image of Lula, as both profiles became a common place for digital informational conversation – both were mentioned by Rousseff’s and Lula’s supporters, as well as by opponents to the two politicians.

Furthermore, in the groups in which Lula’s supporters were concentrated in the fourth context of the empirical analysis, there was the influence of Venezuelan political leadership profiles, such as the country’s president, Nicolás Maduro (@nicolasmaduro) and Diosdado Cabello (@dcabellor), who was, at that time, the President of the National Assembly of Venezuela and also vice-president of PSUV (the party founded by Hugo Chávez); as well as the profiles of leaders who were associated with PT, such as the profile of the party (@ptbrasil) itself, Lula (@lulapelobrasil), Gleisi Hoffmann (@gleisi) – Paraná state’s senator - and Rui Falcão (@rfalcao13) – PT’s national president at that time. In the groups that supported the former president, we also observed the visibility of profiles which belonged to PT’s activism, such as “Brasil da Mudança” (@brasildamundança) and Stanley Burburin (@stanleyburburin); the profiles of the actor and opinion leader José de Abreu (@zehdeabreu) and CUT (@cutnacional), categorized by us as a leadership profile.

These observations about social networks that were constituted on Twitter during the fourth context of analysis require highlighting the strategy of PT’s government communication focused on its militancy (also constituted by the CUT’s and José de Abreu’s profiles) in order to enhance the public image of Lula in face of a favorable result regarding Dilma Rousseff’s impeachment, a strategy that was also used by Venezuelan political leaders. This tactic of building the former president’s public image contributed to the formation of a pro-Lula network that could guarantee greater visibility to his narratives only in the context in which hegemonic media outlets profiles were not central to the circulation of information on Twitter.

In regards to the groups that circulated deconstruction narratives of Lula’s public image, there was a strong influence of hegemonic media outlets (Folha de S. Paulo, O Globo newspaper, Veja magazine, Época magazine), journalistic blogs linked to these outlets and opinion who were opposed to Lula. We believe that due to the reliability of these sites in the circulation of information during the analyzed period the posts published by these profiles became central in the conversations observed in the period and, therefore, the speeches that aimed to deconstruct Lula’s public image were the most visible in these conversations.

It’s important to mention that the centrality of the profiles belonging to the hegemonic media in groups that aimed to detract Lula’s public image reiterates notes regarding the former president’s political journey in which, at various times, the same media had a decisive role in deconstructing or even opposing to Lula’s figure. We can mention the corruption scandals labeled by the press as “mensalão” and “petrolão”, which were disclosed, respectively, in 2005 and 2014. In the case of “mensalão”, Veja and Folha de S. Paulo were the precursors of bad critics/reviews of Lula regarding the alleged involvement of the former president in the scheme. At the “petrolão” scandal, Veja again played a decisive role in presuming Lula’s guilt. As we consider the network analysis of this empirical research, we may state that during the political crisis that occurred between March and May 2016, the major Brazilian media groups continued to play a central role in broadcasting and giving visibility to narratives against Lula.

When it comes to the results of empirical research based on social network analysis, we highlight a third point concerning the inexpressiveness of what we call counter-hegemonic media in the circulation of narratives about Lula in the period of political crisis that occurred between March and May 2016. This counter-hegemonic media is represented by media outlet profiles, such as Carta Capital (@cartacapital), Carta Maior (@cartamaior) and Brasil 247 (@ brasil247), as well as the blog created and managed by Paulo Henrique Amorim, called Conversa Afiada (@conversaafiada). We also emphasize that these profiles outstood inside Lula’s supportive groups in the four analyzed contexts. Their inexpressiveness suggests, consequently, the invisibility of the narratives of Lula’s public image construction published by them.

Given these observations and reflections regarding the social media analysis, we point out that the way the narratives about Lula circulated in the selected period generated a decisive reflection on what was observed in the content analysis stage. Such resonance was perceived, at first, on the terms that were most mentioned in the four investigated contexts, as well as the predominance of themes that aimed to deconstruct the public image of the former president. In these themes, two main narratives were the most visible in the four verified periods: a) the one which aimed to associate Lula with the corruption scandals and to classify him as a manipulative politician responsible for the political crisis in the country - such narrative was the most predominant in the analyzed corpus; and b) the one that aimed to reinforce the end of the former president’s political career in face of the impeachment process suffered by Rousseff. Subsequently, we will analyze these two narratives.

Lula: corrupt and coup-plotter

Regarding the narrative that aimed to associate Lula with an image of a corrupt and a coup-plotter politician, we point out that the recurrence of such allegations was observed based on five specific circumstances: 1) the possibility of Lula’s arrest; 2) the recording and broadcasting of the telephone conversation between Lula and Dilma; 3) the appointment/nomination of Lula to the position of minister of the Civil House; 4) the Operation Car Wash; 5) the social manifestations in favor of Dilma Rousseff’s impeachment. Below, we would like to highlight two examples of this narrative.

1. Veja magazine (@VEJA) “Phone tapping unveils Lula and Dilma’s move against Operation Car Wash, and crowds are back to the streets. https://t.co/uSOQ1ZpYDb”. March 17th, 2016. Tweet.

1. Veja magazine (@VEJA) “Phone tapping unveils Lula and Dilma’s move against Operation Car Wash, and crowds are back to the streets. https://t.co/uSOQ1ZpYDb”. March 17th, 2016. Tweet.

This tweet above exemplified and published on Veja’s profile¹⁷ referred to the title of a story published on March 16^th, 2016 on Veja’s online website and to an image in which Lula appears talking into Dilma’s ears. It should be noted that the magazine used the imagery-visual strategy, as well as the textual one, to deconstruct Lula’s political public image. The article cited in the tweet discusses the content of the conversation between Dilma and Lula, who, according to Veja, revealed Lula’s appointment/nomination as the Minister of the Civil House to avoid the possibility of being arrested for corruption – part of the Operation Car Wash. The photograph used by the magazine to illustrate such news made it possible for the reader of both the tweet and of the journalistic story, to understand the use of the word “coup” as a synonym for plot or cheating, which contributed to the harming of the former president’s image.

The strategy to damage Lula’s public political image, used in the tweet posted by Época magazine and exemplified below, resembles the coup narrative included in the theme about the recording of a telephone conversation between Lula and Dilma (in the tweet published by Veja magazine). However, in Época, the term “coup” was replaced by “plan”, as it can be seen below:

2. Época (@revistaepoca). “EXCLUSIVE: STF ratifies accusation which mentions Dilma and Lula’s plan to spoil Operation Car Wash http://glo.bo/1MEfy8s”. April 18th, 2016. Tweet.

2. Época (@revistaepoca). “EXCLUSIVE: STF ratifies accusation which mentions Dilma and Lula’s plan to spoil Operation Car Wash http://glo.bo/1MEfy8s”. April 18th, 2016. Tweet.

A resemblance between Veja’s and Época›s tweets relies on the strategy of using the visual-imagery narrative to corroborate the text published in the afore mentioned tweet. The image used in Época magazine also shows Dilma talking into Lula’s ears as if both were planning something.

Regarding the use of images in tweets, we might say that, after analyzing the discursive strategies of Veja and Época in the construction of the public image of presidential candidates in the 2016 elections, Silva (2011) points out that both magazines rely on visual content (whether through photographs or illustrations) as one of the main strategies in the deconstruction or construction of candidates. In the specific case analyzed by Silva (2011), the magazines used visual resources to damage Lula’s image, as they also used the same tools to praise the image of his main opponent, Geraldo Alckmin (PSDB). Thus, visual resources, for magazines, play a role of reinforcing, dialoguing and explaining the verbal text.

Therefore, the way images are chosen is based on what the writer intends to say (SILVA, 2011). We emphasize that by including pictures of Lula and Dilma in the aforementioned tweets - in which both are portrayed derogatorily as if they were plotting something - the magazines explored visual content in order to reinforce the critical aspect of the two main characters (SILVA, 2011).

Lula’s “impeachment”

Regarding the narrative of deconstructing Lula’s public image, which aimed to reinforce the end of the former president’s political career alongside the impeachment process suffered by Rousseff, we highlight that an important episode guided the use of this story: the voting of the impeachment in the House of Representatives and the Senate, which occurred respectively in the third and fourth contexts of this analysis. Concerning that, we would like to highlight two examples:

1. Folha de S. Paulo (@folha). “‘I am going home’, says Lula, who was sad after seeing Dilma leaving the Planalto (Executive) http://uol.com/bfjL1P .”. May 12th, 2016. Tweet.

1. Folha de S. Paulo (@folha). “‘I am going home’, says Lula, who was sad after seeing Dilma leaving the Planalto (Executive) http://uol.com/bfjL1P .”. May 12th, 2016. Tweet.

Considering the tweet published by Folha de S. Paulo and shown above, three strategies of deconstruction of Lula’s public image can be observed: 1) one of them was to use the term “sad” as a way of depreciating the mood in which former President Lula was during the voting session in the Federal Senate to open the impeachment process and the likely removal of Dilma; 2) the sentence “I am going home”, which would have been said by Lula at the time of Dilma’s removal from the Presidency. The newspaper’s reproduction of this sentence clearly demonstrates the narrative of deconstruction of the former president’s image by the idea of “impeachment of Lula”, since he would be moving away from Palácio do Planalto (the Executive headquarters in Brasília) alongside Dilma; 3) the third strategy is related to the use of the word “Planalto” as a reference to the place that is no longer inhabited by Lula and Dilma; such tactic is significant in the sense of using the Planalto Palace as a symbol of holding power or a “place of power”. According to the newspaper, the ones who were removed from the “Planalto” - Lula and Dilma – no longer had the attributes of political leadership.

2. Lobão. (@lobaoeletrico). “LULA IS POLITICALLY DEAD. ARREST THIS SCOUNDREL AS SOON AS POSSIBLE”. March 17th, 2016. Tweet.

2. Lobão. (@lobaoeletrico). “LULA IS POLITICALLY DEAD. ARREST THIS SCOUNDREL AS SOON AS POSSIBLE”. March 17th, 2016. Tweet.

As we can see in the tweet above, during the second context of analysis - one day after Dilma’s official announcement about the nomination of Lula for the position of Minister of the House and after the telephone recording between them was leaked, Lobão was clearly against Lula when he referred to the former president as “scoundrel”, also implying that he was an immoral or a dishonest person, as well as demanding his arrest. In the tweet, the influencer seeks to deconstruct Lula’s public image by employing the discursive strategy of ending the former president’s political career by using the phrase “Lula is politically dead”.

Two discursive strategies of Lula’s public image construction - which had lower intensity if compared to the harmful narratives addressed to the former president’s public image - were observed: a) the narrative that saw Dilma’s impeachment as a political and media plot against Lula; and b) the narrative which aimed to praise the social and economic achievements during PT’s presidential terms. In regards to these two angles, we must emphasize the predominance of the strategy of classifying the former president as a victim – that is, someone who needed to be protected from the events that occurred in the analyzed contexts. In the next section, the analysis of the two narratives mentioned above will be presented.

The plot against Lula

The stories told by Twitter users to claim that the impeachment suffered by Dilma Rousseff in 2016 would have also been a coup plotted against Lula was the most predominant among the discursive models of building the former president’s public political image. Here are two examples:

1. PT Brasil. (ptbrasil). “Public acts are scheduled nationwide to defend democracy and the former president Lula https://t.co/DXkIkt8w3z”. March 12th 2016. Tweet.

1. PT Brasil. (ptbrasil). “Public acts are scheduled nationwide to defend democracy and the former president Lula https://t.co/DXkIkt8w3z”. March 12th 2016. Tweet.

As it can be seen in the example above, the tweet was posted by PT’s official Twitter account during the first analysis scenario. Although in the first field of analysis there were social demonstrations in favor of Dilma’s impeachment, this tweet mobilized users for public acts that would take place during the second context of analysis (the protests against the coup on March 18^th, 2016).

At the end of this tweet, the URL “https://t.co/DXkIkt8w3z” directed the reader to a Facebook page related to the March 18^th’s demonstration summoning event entitled “On the streets against the coup and in defense of democracy”. The event was created by Frente Brasil Popular, which was one of the main groups that articulated the demonstrations against the coup in 2016.

The verb “defend” – used on PT’s Twitter account - conveyed the idea that both former President Lula and the Brazilian democracy were victims of a coup, and that they needed to be protected - in the case of this tweet, the form of protection or defense was through the social demonstrations that took place on March 18^th, 2016. We also observe the party’s strategy of centralizing the circumstance of the political coup, not on Dilma’s figure, but on Lula’s, which demonstrates the protagonism of the former president’s public representation during the analyzed period.

2. Maduro, Nicolás. (NicolasMaduro). “We will raise the voice of global solidarity to Dilma and Lula against this media-judicial Coup in Brazil…”. March 17th, 2016. Tweet18

2. Maduro, Nicolás. (NicolasMaduro). “We will raise the voice of global solidarity to Dilma and Lula against this media-judicial Coup in Brazil…”. March 17th, 2016. Tweet18

As it can be seen in the tweet above, Nicolás Maduro, Venezuelan president at the time, used the term “media-judicial coup” to refer to Dilma Rousseff’s impeachment lawsuit and to claim that such process would be a plot engineered by both the media and the Brazilian Justice. We emphasize that, by using the expression “media coup”, Maduro resorted to narratives commonly attributed to speeches given by Hugo Chavez, former president of Venezuela and one of the main leaders of the Latin American left-wing. In such pronouncements, Chávez repeatedly blamed Venezuela’s private media for violating their ideals of government (PASSOS, 2012).

Maduro also used the mobilization narrative to push Twitter users to sympathize with Dilma and Lula against the coup. This narrative is justified by the fact that the tweet was published on March 17^th, 2016, the day before the social demonstrations against the impeachment.

Lula’s legacy

The narrative regarding Lula’s legacy was the second most predominant strategy for building the public image of the former president in tweets published in the four contexts of analysis. Regarding the recurrence of the former president’s legacy story, two important points must be highlighted: 1) this was the least expressive narrative on Twitter during the analyzed period, compared to all the other previously analyzed narratives - both in the construction and deconstruction of the former president’s public image; 2) the narrative about Lula’s legacy was identified mainly when the term “poverty” was mentioned - this word was one of the most cited in the third context of analysis, in which the House of Representatives voted to carry on with the impeachment process. There are two examples of this narrative below:

1. Foresi, Dante L. (DanteForesi). “Lula and Dilma’s terms freed 40 million of Brazilians from poverty, where the ones who vote for the impeachment today were in at that time.” April 17th, 2016. Tweet19. 2. Internacionales CTA. (RRIICTA). “We, workers, will never forget that Lula and Dilma freed 40 million Brazilian brothers and sisters from poverty”. April 18th, 2016. Tweet20.

1. Foresi, Dante L. (DanteForesi). “Lula and Dilma’s terms freed 40 million of Brazilians from poverty, where the ones who vote for the impeachment today were in at that time.” April 17th, 2016. Tweet19.

2. Internacionales CTA. (RRIICTA). “We, workers, will never forget that Lula and Dilma freed 40 million Brazilian brothers and sisters from poverty”. April 18th, 2016. Tweet20.

It is possible to observe, considering the examples above, that the narrative of Lula’s public image construction, based on his political heritage, was similarly used by the three types of profiles (political leaders, opinion leaders and regular users). The central idea of this discursive strategy was that millions of Brazilians would have left poverty due to the efficiency of Lula’s and Dilma’s governments. In addition to that, the tweets published by journalist Dante Foresi (@DanteForesi) and by the Secretariat of International Relations of the Central Workers of Argentina (@RRIICTA) enabled us to quantify the number of Brazilians who would have come out of poverty due to PT’s ruling era. Thus, the content of these tweets included the number “40 million” to validate the importance of such legacy.

Final considerations

Observing the narratives on Twitter – which are conflicting – in the four contexts of the empirical analysis, we conclude that there was a polarization of meanings from a dispute between Twitter users who published narratives aiming to blame the former president for the situation in the country as opposed to the profiles that referred to him as a victim of the same situation. Thus, a friend versus enemy clash was observed (LATTMAN-WELTMAN, 2015, SOLANO; ORTELLADO; MORETTO, 2017), which was established between Lula’s supporters, represented by Venezuelan and PT political leaders, as well as the Lulist militancy, and the opponents of the former president, associated with the hegemonic media and bloggers linked to these outlets.

It is necessary to build a parallel between the results of our empirical research and the data regarding the 2018 Presidential election voting intentions conducted by the DataFolha Institute and released on April 11^th, 2016, in which Lula was appointed as the presidential candidate of the country by respondents. However, the same study measured the electoral rejection rate of the names included in the scenarios presented and, despite the favorable result for Lula, the former president’s name was the one with the highest rejection rate by respondents (DATAFOLHA, 2016). Therefore, it is noted that this dubiety of the Brazilian electorate was reflected in the conversations about Lula in the four contexts.

We also point out that the narrative clash observed on Twitter in the contexts of empirical analysis occurred incisively from the intense battle over the meaning of the term “coup”, which was used both in the narrative in which Lula was presented as guilty by the Brazilian political crisis and in the stories that portrayed him as a victim of such facts.

Thus, regarding the dispute over the meaning of the word “coup”, it is appropriate to reflect on the strategy used by the hegemonic outlets in the analyzed period to redefine the meaning of this narrative, which was addressed and reinforced by opponents of Dilma Rousseff’s impeachment lawsuit. The legitimizing groups of the impeachment narrative as a coup accused the hegemonic media of being one of the main articulators of the plot against Brazilian democracy. According to this perspective, therefore, the media would be a scammer/plotter. But what was noticed in the empirical research was that the same media strategy was used to reverse such accusation, that is, Lula became responsible for the coup instead of the communication outlets. This strategy was also noted in narratives published by some aforementioned opinion leaders (Lobão and Danilo Gentili) as well as regular users, who mostly portrayed themselves on the microblog as Antilulist activists.

Due to the credibility given to the profiles of hegemonic media when it comes to the circulation of information in the analyzed period, the narrative that framed Lula as a plotter was the one that generated greater awareness amongst the four contexts of empirical analysis, which contributed to the inexpressiveness of the story that the former president would be the victim of a coup.

Such observation seems to be crucial to the investigation carried out in the present research since there were many narratives in circulation in the analyzed period. However, it should be noted that the power of hegemonic outlets – when adopting the same attitude toward the former president - was possible due to their ability to make certain narratives more visible than others. Therefore, there was a clash between the pro-Lula network, consisting of government communication, Latin American leaders and the PT militancy, and the opposing group, composed of hegemonic media outlets, bloggers linked to these channels, opinion leaders and regular users who were against the former president. In this dispute for visibility, the referability and reliability acquired by the opponents made the narratives of deconstruction of Lula’s public image more evident. This observation is not so different from what was observed in the former president’s political journey, in which the media played a core role in destabilizing Lula’s reputation in the face of the Brazilian collective imagination, which shows that the circulation of information online, despite offering a “place” for discussion and civil conversations, reinforces informational hierarchies already established offline.

• 1
  This article is a result of a PhD research which took place in 2013 and 2017, see Passos (2017).
• 2
  Lula won the 2002 presidential election on the runoff with 61,27% of all votes (nearly 52 million votes); the second-placed candidate, José Serra, from Partido da Social Democracia Brasileira (PSDB), had 38,72% (nearly 33 million votes), according to data provided by the Tribunal Superior Eleitoral (Superior Electoral Court). Available at: http://www.tse.jus.br/eleicoes/eleicoes-anteriores/eleicoes-2002/resultados-das-eleicoes-2002. Accessed on: 1 jan. 2016.
• 3
  According to Ricci (2010), the Lulism was born as an attempt to manage the State and the Brazilian governability and it has Lula as its main character.
• 4
  Information obtained from the Tribunal Superior Eleitoral (Superior Electoral Court). Available at: http://www.tse.jus.br/eleicoes/eleicoes-anteriores/eleicoes-2006/resultado-da-eleicao-2006. Accessed on: 1 jan. 2016.
• 5
  Regarding this prestige transference from Lula to Dilma, during the 2010 presidential elections, we recommend the reading of Telles e Pires (2015).
• 6
  Dilma won the 2010 presidential election in the second round with 56.05% of the vote (about 55 million votes) against opposition candidate José Serra (PSDB) who won 43.95% of the vote (about 43 million votes) according to data from the Superior Electoral Court. Available at: http://www.tse.jus.br/eleicoes/eleicoes-anteriores/eleicoes-1998/resultados-das-eleicoes-de-2010. Accessed on: 1 jan. 2016
• 7
  A hashtag is a context label on Twitter, which points out to a specific term that builds a context, as it also enables users to search for those particular tweets. In general, it is presented with the graphic sing of
• 8
  Dilma won the 2014 elections on the runoff with 51,64% of valid votes, against Aécio Neves (PSDB) who had 48,36% of the votes – it was the toughest election since Brazil’s redemocratization process (AMARAL; RIBEIRO, 2015).
• 9
  The first of the three scenarios, in which Lula and Marina Silva were racing for the voting intention leadership also included the presidential candidate Aécio Neves (PSDB party) – at that time, he was also the president of the party, a senator representing Minas Gerais state and the strongest competitor to Dilma Rousseff during the 2014 presidential elections; the second scenario had Geraldo Alckmin (PSDB) as a competitor – at that moment, he was São Paulo state’s governor and also Lula’s competitor in the 2006 elections; the third scenario included José Serra (PSDB) – at that time, Serra was a senator representing São Paulo state, a competitor to Dilma Rousseff in the 2010 elections and also a competitor to Lula in the 2002 elections. The fourth possibility, in which Lula led the intention of votes, brought Sérgio Moro’s name, who had no party so far and who was also the federal judge responsible for the Operation Car Wash.
• 10
  This tool, which is available online (www.netlytic.com), is labeled by its developers as “a cloud-based social media and text analytical tool which is able to automatically summarize and discover social media chats in social network websites. It uses public APIs in order to collect posts from Twitter, YouTube and Facebook (public pages) and supports the analysis of its own data collection (GRUZD, 2017). This research used the paid-version of the software.
• 11
  It is important to state that Becker et al (2016), in an article called “Manifestações e votos ao impeachment de Dilma Rousseff na primeira página de jornais brasileiros” (Protests and votes for Dilma Rousseff’s impeachment on the Brazilian newspapers’ frontpages), published in 2016, also elected the same contexts: the protests on March 13^th and March 18^th as well as the impeachment voting sessions, at the Deputy Chamber/House of Representatives and at the Senate, respectively on April 17^th 2016 and on May 12^th 2016, as a background to analyze how the newspapers O Globo, Folha de S. Paulo and O Estado de S. Paulo, covered this aforementioned facts. It’s important to emphasize that these choice for these authors endorse the relevance of the four contexts in the Brazilian contemporary political conjuncture and it also ensures that the chosen methodology is right for this research.
• 12
  The decision to collect data related to dates before the protests is due to the fact that we aim to understand the published messages during these events as well as before them. The dates related to the period after the impeachment voting sessions were chosen as we aim to understand the content of such posts during the parliamentarian decisive lawsuit and also its repercussion after the results of such voting sessions.
• 13
  These categories were built based on a concept called “types of player” (or Twitter users’ types), coined by Moraes e Quadros (2016).
• 14
  Gephi is described by Bastian, Hymann e Jacomy (2009) as an open-code software for graphic producing and network analyzing. It uses a three-dimensional rendering mechanism in order to showcase great networks in real time, as well as speeding their exploration. Its flexible and multitask interface brings new possibilities in working with complex data and producing valuable visual results which (BASTIAN; HEYMANN; JACOMY, 2009).
• 15
  Available at: http://www.laurenceanthony.net/software.html. Accessed on: 1 jan. 2016.
• 16
  As this article is a result of a PhD research, it was necessary to reduce the number of terms displayed here from 80 to 20 in each of the analyzed contexts.
• 17
  The tweet is also composed by a URL (https://t.co/uSOQ1ZpYDb) which makes a connection between the message posted on Twitter and the Veja news website, in which it’s possible to read the full article.
• 18
  Maduro, Nicolás (NicolasMaduro). “Levantemos la Voz de la Solidaridad Mundial con Dilma y Lula ante el Golpe Mediático-Judicial en Brasil...”. 17 mar. 2016. Tweet.
• 19
  Foresi, Dante L. (@DanteForesi). “Los gobiernos de Lula y Dilma sacaron a 40 millones de brasileños de la pobreza donde los habían metido los que hoy votan por el impeachment.”. April 17th 2016. Tweet.
• 20
  Internacionales CTA. (@RRIICTA). “Los trabajadores nunca olvidaremos que Lula y Dilma sacaron de la pobreza a 40 millones de hermanos brasileros.”. April 18th 2016. Tweet.

Referências

• AMARAL, O. E.; RIBEIRO, P. F. Por que Dilma de novo? Uma análise exploratória do Estudo Eleitoral Brasileiro de 2014. Revista de Sociologia e Política, Curitiba, v. 23, n. 56, p. 107-123, dez. 2015.
• BARDIN, L. Análise de conteúdo Lisboa: Edições 70, 1977.
• BASTIAN, M.; HEYMANN, S.; JACOMY, M. Gephi: an open source software for exploring and manipulating networks. In: INTERNATIONAL ICWSM CONFERENCE, 3., 2009. Proceedings… p.361-362. Disponível em: http://www.aaai.org/ocs/index.php/ICWSM/09?paper/download/154/1009 Acesso em: 15 mar. 2015.
  » http://www.aaai.org/ocs/index.php/ICWSM/09?paper/download/154/1009
• BAUER, M. W. Análise de conteúdo clássica: uma revisão. In: BAUER, M. W.; GASKELL, G. (Ed.). Pesquisa qualitativa com texto, imagem e som: um manual prático. Petrópolis: Vozes, 2008. p. 189-217.
• BECKER, C. et al. Manifestações e votos ao impeachment de Dilma Rousseff na primeira página de jornais brasileiros. Revista Latinoamericana de Ciencias de la Comunicación, v. 13, n. 24, jun. 2016. Disponível em: http://www.alaic.org/revistaalaic/index.php/alaic/article/view/736 Acesso em: 15 jun. 2017.
  » http://www.alaic.org/revistaalaic/index.php/alaic/article/view/736
• CONFEDERAÇÃO NACIONAL DA INDÚSTRIA. Pesquisa CNI – IBOPE: avaliação do governo e expectativa em relação ao novo governo. Brasília, dez. 2010. Disponível em: http://www.ibopeinteligencia.com/inteligencia/downloads/2010/JOB_192-19_BRASIL_-_Relatorio_de_tabelas_(imprensa).pdf/ Acesso em: 15 mar. 2016.
  » http://www.ibopeinteligencia.com/inteligencia/downloads/2010/JOB_192-19_BRASIL_-_Relatorio_de_tabelas_(imprensa).pdf/
• DATAFOLHA. Intenção de voto para presidente da República: PO813859 07 e 08/04/2016. Datafolha, São Paulo, 11 abr. 2016. Disponível em: http://media.folha.uol.com.br/datafolha/2016/04/11/intencao_de_voto_presidente.pdf Acesso em: 15 mar. 2017.
  » http://media.folha.uol.com.br/datafolha/2016/04/11/intencao_de_voto_presidente.pdf
• FRAGOSO, S.; RECUERO, R.; AMARAL, A. Métodos de pesquisa para internet Porto Alegre: Sulina, 2013.
• GLASENAPP, R. B.; FARIAS, A. O processo de impeachment na Constituição Federal de 1988 e sua aplicação no caso Dilma Rousseff. Revista Paradigma, Ribeirão Preto, v. 25, n. 1, p. 79-101, jan. /jun. 2016.
• GREIMAS, A. J. Semântica estrutural São Paulo: Cultrix/EdUSP,1973.
• GRUZD, A. Research with Netlytic Netlytic: Software for Automated Text and Social Network Analysis. 2017. Disponível em: https://netlytic.org/home/?page_id=49 Acesso em: 15 mar. 2015.
  » https://netlytic.org/home/?page_id=49
• LATTMAN-WELTMAN, F. Democracia e revolução tecnológica em tempos de cólera: influência política e midiática e radicalização militante. In: CONGRESSO DA ASSOCIAÇÃO BRASILEIRA DE PESQUISADORES EM COMUNICAÇÃO E POLÍTICA, 6, 2015, Rio de Janeiro. Anais... Rio de Janeiro: Compolítica, 2015. p.1-22.
• MALINI, F. et al. #VemPraRua: narrativas da revolta brasileira. 2014. Disponível em: http://labic.net/wp-content/uploads/VemPraRua-Narrativas-da-Revolta-brasileira.pdf Acesso em: 20 mar. 2016.
  » http://labic.net/wp-content/uploads/VemPraRua-Narrativas-da-Revolta-brasileira.pdf
• MORAES, T. P. B.; QUADROS, D. G. A crise do governo Dilma Rousseff em 140 caracteres no Twitter: do #Impeachment ao #ForaDilma. Em Debate, Belo Horizonte, v. 8, p. 14-41, 2016.
• PASSOS, M. R. Mito e narrativa: a (des)construção da imagem pública de Lula no contexto da crise política de 2016. 2017. 326f. Tese (Doutorado) – Universidade de São Paulo, Programa de Pós-Graduação em Ciências da Comunicação.
• PASSOS, M. R. O Twitter como trincheira de luta: um estudo sobre o microblog de Hugo Chávez. 2012. 111 f. Dissertação (Mestrado) – Pontifícia Universidade Católica de Minas Gerais, Programa de Pós-Graduação em Comunicação Social.
• RECUERO, R. “Deu no Twitter, alguém confirma?” Funções do jornalismo na era das redes sociais. In: ENCONTRO NACIONAL DE PESQUISADORES EM JORNALISMO, 9, 2011, Rio de Janeiro. Rio de Janeiro: ECO – Universidade Federal do Rio de Janeiro, 2011. Disponível em: http://www.raquelrecuero.com/arquivos/sbpjorrecuero.pdf Acesso em: 15 ago. 2016.
  » http://www.raquelrecuero.com/arquivos/sbpjorrecuero.pdf
• RECUERO, R. #ProtestosBR: análise comparativa do discurso dos veículos jornalísticos. In: ENCONTRO NACIONAL DE PESQUISADORES EM JORNALISMO, 12, 2014, Santa Cruz do Sul. Anais... Brasília: SBPJor, 2014, p.1-20.
• RECUERO, R.; ZAGO, G. A Economia do Retweet: Redes, Difusão de Informações e Capital Social no Twitter. Revista Contracampo, Niterói, v. 24, n. 1, p. 19-43, jul. 2012.
• RECUERO, R.; ZAGO, G.; BASTOS, M. Análise de redes para mídia social Porto Alegre: Sulina, 2015.
• RICCI, R. Lulismo. Da era dos Movimentos Sociais à Ascensão da Nova Classe Média Brasileira Brasília: Fundação Astrogildo Pereira, 2010.
• SANTAELLA, L.; LEMOS, R. Redes sociais digitais: a cognição conectiva do Twitter. São Paulo: Paulus, 2010.
• SCARTEZINI, N. A fascistização da indignação: as manifestações de 2015 no Brasil. Cadernos de Campo, Araraquara, n. 20, p. 183-206, jan./jun. 2016.
• SILVA, J. C. R. Época, Veja e o e(leitor): estratégias discursivas na construção da imagem de presidenciáveis. 2011. 206f. Tese (Doutorado em Linguística) – Instituto de Letras, Universidade de Brasília, Brasília, 2011.
• SOLANO, E.; ORTELLADO, P.; MORETTO, M. 2016: o ano da polarização? Friedrich Ebert Stifung Brasil, Análise, n.22, 2017.
• TELLES, H.; PIRES, T. M. C. C. A comunicação política para o “terceiro mandato” de Lula: a transferência de prestígio nas eleições presidenciais brasileiras de 2010. In: TELLES, H.; LAVAREDA, A. (Orgs.). Voto e estratégias de comunicação política na América Latina Curitiba: Appris, 2015. p.121-153.

Publication Dates

• Publication in this collection
  2 Dec 2019
• Date of issue
  Sep-Dec 2019

History

• Received
  22 Feb 2019
• Accepted
  12 Sept 2019