Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
110 tokens/sec
GPT-4o
56 tokens/sec
Gemini 2.5 Pro Pro
44 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Hate Lingo: A Target-based Linguistic Analysis of Hate Speech in Social Media (1804.04257v1)

Published 11 Apr 2018 in cs.CL and cs.SI

Abstract: While social media empowers freedom of expression and individual voices, it also enables anti-social behavior, online harassment, cyberbullying, and hate speech. In this paper, we deepen our understanding of online hate speech by focusing on a largely neglected but crucial aspect of hate speech -- its target: either "directed" towards a specific person or entity, or "generalized" towards a group of people sharing a common protected characteristic. We perform the first linguistic and psycholinguistic analysis of these two forms of hate speech and reveal the presence of interesting markers that distinguish these types of hate speech. Our analysis reveals that Directed hate speech, in addition to being more personal and directed, is more informal, angrier, and often explicitly attacks the target (via name calling) with fewer analytic words and more words suggesting authority and influence. Generalized hate speech, on the other hand, is dominated by religious hate, is characterized by the use of lethal words such as murder, exterminate, and kill; and quantity words such as million and many. Altogether, our work provides a data-driven analysis of the nuances of online-hate speech that enables not only a deepened understanding of hate speech and its social implications but also its detection.

User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (5)
  1. Mai ElSherief (14 papers)
  2. Vivek Kulkarni (33 papers)
  3. Dana Nguyen (2 papers)
  4. William Yang Wang (254 papers)
  5. Elizabeth Belding (18 papers)
Citations (278)

Summary

Analysis of Hate Speech Targeting on Social Media

The exponential growth of social media has not only democratized communication and enhanced connectivity but also facilitated a rise in antisocial behavior, notably online hate speech. The paper "Hate Lingo: A Target-based Linguistic Analysis of Hate Speech in Social Media" by ElSherief et al. provides a meticulous investigation into two primary categories of hate speech differentiated by their target: Directed hate speech aimed at individuals or entities, and Generalized hate speech targeting groups unified by a protected characteristic. This work extends beyond the prevalent binary hate/non-hate classification and offers a comprehensive lens into the nuanced forms of hate speech.

Methodology and Dataset

Using a sizable dataset derived from various sources, including Twitter Streaming API samples and external hate speech corpora, the authors meticulously categorize tweets into Directed and Generalized hate speech. Importantly, they deploy linguistic and psycholinguistic analyses, facilitated by tools like LIWC and SemaFor, to identify and contrast the patterns characteristic of each form of hate speech. This data-driven approach uncovers significant linguistic markers that distinguish Directed from Generalized hate speech.

Characteristic Features

Directed hate speech is invariably more personal and direct, characterized by informality, high levels of anger, and language that explicitly attacks the target. The presence of personal pronouns and second-person directives underscores its hostile interpersonal nature. Psycholinguistic analysis reveals that Directed hate speech often reflects lower levels of analytical thinking and a higher sense of clout and dominance.

Conversely, Generalized hate speech frequently utilizes religious and ethnic terms and is marked by broader and more abstract constructs, as evidenced by the prevalence of quantity words and lethal terms. Generalized hate speech is generally less informal and is especially characterized by an "us versus them" mentality substantiated by the frequent use of third-person pronouns.

Implications

The distinction elucidated in this paper has significant legal and societal implications. The differentiation between Directed and Generalized hate speech impacts how policies around free speech may be framed and understood, particularly concerning targets' legal rights. Directed hate speech can lead to more immediate emotional harm, while Generalized hate speech can influence public opinion more broadly, emphasizing the complexity in designing interventions for each type.

Contributions and Future Directions

The paper contributes significantly by pioneering an in-depth, target-based analysis of hate speech and providing a detailed dataset that is now available for public and academic use. The findings call for nuanced approaches in designing more effective automated hate speech detection systems, proposing a departure from traditional methods that overlook the complexity of target-based nuances. Furthermore, the paper paves the way for exploring counter-speech interventions, urging a focus not only on language but also the societal roles and motivations behind hate speech participation.

In summary, ElSherief et al. advance our understanding of hate speech dynamics on social media through a granular analysis informed by who is targeted and how. Their work lays foundational stones for ongoing research and initiatives aimed at addressing and mitigating the impact of online hate speech in society. The efficacy of future detection and intervention strategies hinges upon such fine-grained insights, steering towards a digitally harmonious environment.