During the period of the 2018’s presidential election in Brazil a movement of polarization was observed, especially during the second round, which concentrated in the opposition between the groups that supported the candidates Jair Bolsonaro (PSL) and Fernando Haddad (PT). In this way this article aims to approach in which ways the polarized groups built representations around the opposite group and established a relationship with the (O)other in the discursive space of Twitter, since it was a space of wide spreading and (re)production of the polarized discourses. Is presented as corpus excerpts from posts that circulated on Twitter from October 7th to 27th, 2018. Besides, the Discourse Analysis, especially through discourse studies in French perspective, is taken as theoretical and analytical device, in order no work with the language and imagetic materiality of the utterances. The qualitative and interdisciplinary study seeks support in the methodological perspective of digital ethnography, in the way of conceiving and studying the polarized communities that surround the social media. The groups called as “communists” and “bolsominions”, according to the representations created in this period, established a relation of harassment with the (O)other, presenting sayings that offended, delegitimized and, in many moments, removed the (O)other of their position as a subject. The defense of an “absolute truth” by the groups contributed for the emergency of a discourse violence, using swearwords and ad hominem attacks. During the 2018’s election the (O)other was put as an “outsider”, taking, in many moments, the position as a “strange”, “abominable”.

The objective of this dissertation was to analyze and describe the violent discourse manifested in the form of an online commentary in publications of the Folha de São Paulo newspaper in fanpage on Facebook. This objective is based on the concepts society, discourse and cognition coming from the perspective of the cognitive theory of strategic information processing (VAN DIJK 1983b, 1984a, 2004, 2005, 2006, 2008, 2011, 2016a and 2016b), sociocognitive discourse theory (HAMILTON, 1981a; TAJFEL, 1981, 1982) and the theory of social cognition (ARGYLE, FURNHAM, & GRAHAM, 1981; FORGAS, 1979, 1981; FURNHAM & ARGYLE, 1982; VAN DIJK, 1983). The methodology that guided the achievement of the above mentioned objective is of a qualitative and interpretative nature, being based on an ethnographic case study approach, because we can describe the culture involved in the production of hate speech that has gained space in the virtual environment, thus, this research aims to describe the behavior of social actors and the meanings that the online enunciators attribute to discursive acts and virtual interactions (SPRADLEY, 1979). By means of this methodological aim, we generated a corpus of 696 online comments, that were collected at Folha de São Paulo's Facebook Fanpage, which were analyzed through the techniques of the Linguistic-Sociocognitive Model (VAN DIJK, 2016a) and the Systems of Beliefs and Social Mental Models (VAN DIJK, 2016b). The results found allow us to reach the conclusion that the semantic and cognitive models constructed by the technosocial actors in the posts of Folha de São Paulo act expressively and/or significantly in the orientation of the formulation of online comments. Based on these findings, we can state that the belief systems and context conditions guide both the positioning and the orientation of the construction of the senses in online comments.