Targets of Online Hate Speech in Context

Abstract

Online hate speech, especially on social media platforms, is the subject of both policy and political debate in Europe and globally - from the fragmentation of network publics to echo chambers and bubble phenomena, from networked outrage to networked populism, from trolls and bullies to propaganda and non-linear cyberwarfare. Both researchers and Facebook Community standards see the identification of the potential targets of hateful or antagonistic speech as key to classifying and distinguishing the latter from arguments that represent political viewpoints protected by freedom of expression rights. This research is an exploratory analysis of mentions of targets of hate speech in comments in the context of 106 public Facebook pages in Romanian and Hungarian from January 2015 to December 2017. A total of 1.8 million comments were collected through API interrogation and analyzed using a text-mining niche-dictionaries approach and co-occurrence analysis to reveal connections to events on the media and political agenda and discursive patterns. Findings indicate that in both countries the most prominent targets mentioned are connected to current events on the political and media agenda, that targets are most frequently mentioned in contexts created by politicians and news media, and that discursive patterns in both countries involve the proliferation of similar stereotypes about certain target groups.

Swedish journalists' perceptions of legal protection against unlawful online harassment

“…As such, these results also confirm and demonstrate the effects of what previous studies have already described; namely how the collective ambition to demonize professional journalism effectively suppresses journalistic voices (cf. Barlow and Awan, 2016 ; Meza et al, 2018 ; Sobieraj et al, 2020 ; Carlson et al, 2021 ; Posetti et al, 2021 ). Furthermore, this study indicates that the legal system is inadequate in countering the adverse control that online harassment collectively constitutes.…”

“…Instead, it is a speech issue with substantial implications for journalism (Waisbord, 2020 ). Several other scholars support this claim by connecting the aim of populist movements to control of public discourse by attacking facts, news, and information with the increased demonization of journalism (Barlow and Awan, 2016 ; Meza et al, 2018 ; Sobieraj et al, 2020 ; Carlson et al, 2021 ; Posetti et al, 2021 ).…”

Discurso de ódio nas redes sociais digitais: tipos e formas de intolerância na página oficial de Jair Bolsonaro no Facebook

“…No caso desta pesquisa, o interesse está no tipo e na forma do discurso de ódio manifestado em canais digitais administrados por figuras políticas. O estudo leva em consideração que o assédio nas redes sociais digitais, em muitos casos, aparece como comunicação estratégica, na qual mensagens específicas são criadas por atores políticos para manipular a opinião pública (MEZA et al, 2018) e "degradar simbolicamente grupos historicamente oprimidos ou sistematicamente discriminados" (SPONHOLZ, 2020, p. 221).…”

Offensive Text Span Detection in Romanian Comments Using Large Language Models

“…In the context of Romanian social media, Meza et al [8] highlighted that the mostprevalent target groups for offensive messages are welfare recipients or poor people, followed by the Roma and Hungarian communities. This finding supported the 2018 study [9] by the Elie Wiesel National Institute for Studying the Holocaust in Romania.…”