Definition generation models trained on dictionary data are generally expected to produce neutral and unbiased output while capturing the contextual nuances. However, previous studies have shown that generated definitions can inherit biases from both the underlying models and the input context. This paper examines the extent to which stance-related bias in argumentative data influences the generated definitions. In particular, we train a model on a slang-based dictionary to explore the feasibility of generating persuasive definitions that concisely reflect opposing parties’ understandings of contested terms. Through this study, we provide new insights into bias propagation in definition generation and its implications for definition generation applications and argument mining.

While persuasion has been extensively examined in the context of politicians’ speeches, there exists a notable gap in the understanding of the pathos role in user-generated argumentation. This paper presents an exploratory study into the pathos dimension of user-generated arguments and formulates ideas on how pathos could be incorporated in argument mining. Using existing sentiment and emotion detection tools, this research aims to obtain insights into the role of emotion in argumentative public discussion on controversial topics, explores the connection between sentiment and stance, and detects frequent emotion-related words for a given topic.