A LINGUISTIC ANALYSIS OF CAUSALITY IN HATE SPEECH

This dissertation investigates how linguistic phenomena are used to manipulate the perception of causal responsibility and blame when expressing hate speech. The research questions addressed are as follows: (1) How is responsibility being assigned? (2) Are different types of causal constructions and strategies used similarly in hate speech? (3) Do the strategies vary based on language or are they universal? (4) How significant is causal responsibility in hate speech? To answer research questions (1) and (2), I conducted a corpus analysis of select causal constructions and strategies of indirect speech using annotated English hate speech datasets to see what syntactic choices were made in hate speech when referencing minority (targeted) groups and majority groups. I found that hate speech tends to use causal constructions to minimize the causal responsibility of majority groups and indirect speech strategies such as word order and passivization to focus on minority groups' involvement in events. To answer research question (3), I extended the corpus study to include English, Spanish, German, and French. The same causal constructions and strategies of indirect speech were analyzed. I found that the strategies used to manipulate the assignment of causal responsibility do vary cross-linguistically. For example, passive voice and causative verbs were used more frequently in English hate speech and implicit causality verbs were used more in Spanish and German hate speech. However, none of the syntactic choices examined were used to assign less causal responsibility to minority groups than majority groups in hate speech. Finally, to answer research question (4), I trained models to detect hate speech using causal information to test how predictive and generalizable the patterns regarding the assignment of causal responsibility in hate speech are. I found that the models did learn patterns related to the assignment of causal responsibility and that those patterns generalize better than patterns related to lexical choice.

Name thesis:degree_name

Ph. D., Linguistics

Level thesis:degree_level

Doctor of Philosophy

Discipline thesis:degree_discipline

Linguistics

Grantor

Cornell University

Year dc:date.issued

2024

Author dc:creator

• Lamp, Kaelyn

Committee members dc:contributor.committeemember

• Aparicio Terrasa, Helena
• Lyon, Mary
• Murray, Sarah

• Causality
• Hate Speech
• Linguitics

Statement dc:rights

• Attribution 4.0 International

Licence dc:rights.uri

• https://creativecommons.org/licenses/by/4.0/

Language dc:language.iso

en

DOI

https://doi.org/10.7298/rws8-f777

Dc Identifier Other

ProQuest Submission ID: 14603
ProQuest Publication ID: 31489093

OAI identifier oai:identifier

oai:ecommons.cornell.edu:1813/116496