Project Description

Social media channels have come to play an increasingly large role in our everyday lives and communication. They provide a context in which people across the world can communicate, share knowledge, exchange messages, and interact with each other irrespective of the distance and the social differences between them, thus allowing greater freedom of expression and empowering individual voices. At the same time, however, these social media channels also enable anti-social behaviour, cyberbullying, online aggression, and hate speech. In this new communication culture, which has been adopted and adapted in other contexts such as politics and other forms of public speech, appeals to emotions and personal beliefs are important persuasive devices. Verbal aggressions, offensive propaganda, and the construction of authority and subordination both in speech and in writing have become frequent traits and instruments of communication in various spheres of our society.

Language is a key element in the construction and reinforcement of social identities, and as a consequence, also in the creation and diffusion of stereotypes, discrimination, and social injustices. The use of language to attack an individual or group based on attributes such as race, ethnicity, religion, gender, nationality, political ideology, disability or sexual orientation constitutes the basis of what is now known as hate speech. In this project, researchers will investigate and discuss hate speech from a linguistic perspective. The main objectives are:

1) IDENTIFICATION: to identify the common linguistic characteristics and features of hate speech in different domains of communication, focusing on Italian as an empirical testbed;

2) APPLICATION: to produce an annotated catalogue of these features that could serve as a guideline for the recognition (with the additional help of natural language processing methods) and avoidance of hate speech, as well as for the distinction between hate speech and freedom of expression;

3) EXTENSION: to compare and contrast our findings with datasets from other European languages, so as to assess the viability of extending our results to other languages and to investigate the possible distinction between general vs language-specific features of hate speech.

Impact

The linguistic contribution to the identification of hate speech will have a wide domain of application and a broad social impact. The response to hate speech should not be the limitation or repression of free speech; instead, social media users and people exposed to public speeches should be educated in the recognition and hence avoidance of hate speech. Our results will therefore make a valuable contribution in all areas of communication that may be affected by hate speech.

First, our findings can be used as teaching materials in University courses on communication, pragmatics, text analyses, and general linguistics, particularly as part of degrees in Italian Studies. We know from experience that it is not only native speakers, but also foreign language students who find these issues particularly attractive and interesting. As stated above, the material will initially look at Italian, but will then be adapted to other languages.

Second, our guidelines will also appeal to all journalists, writers, and public figures who, for personal or professional reasons, are trying to avoid hate speech in their texts or speeches and want to communicate in a more polite way and at a more appropriate level. They will also increase awareness among all users of social media, helping them to communicate in a just, fair, and non-offensive way. We also envisage that individuals or institutions engaged in legal studies might be interested in our work, insofar as our guidelines could help legislators, judges and lawyers to arrive at a more precise definition of hate speech and offensive language. Several countries either make no mention of hate speech in their current legislation or use broad and ambiguous definitions and applications. A linguistic definition of hate speech that goes beyond the more straightforward and overt use of offences and insults could help us to understand what is illegal and what is not.

These areas of potential application and social impact will ultimately contribute to the Sustainable Development Goals of quality of education, equality, peace, justice, and strong institutions, and to the defence and protection of human rights. Indeed, education, instruction, and equality in this respect are only possible when hate speech is easily and concretely identifiable and recognisable.

Once goals (1) and (2) are achieved, we will approach educational institutions and/or private foundations in search of external funding to further develop our project. To this end, we will also explore the possibilities of collaborating with computational linguists who work on natural language processing and automatic detection of hate speech (along the lines of work such as Schmidt & Wiegand 2017 and Fortuna & Nunes 2018). To this end, we will consult the members of an on-going project on automatic meaning detection at the KU University, who are also members of the proposed project (i.e. Stefania Marzo and Stefano De Pascale), especially with respect to the most appropriate methods for retrieving semantic information from corpora.

Further Reading

See the recent Council of Europe work on Hate Speech, which includes specific recommendations on combating hate speech.