The unbearable hurtfulness of sarcasm

作者：

Highlights：

• Hurtful language is consistently found in a corpus of Italian sarcastic language.

• Error analysis shows that hostility triggers ironic and sarcastic interpretation.

• Negative emotion features improve irony detection with a neural language model.

• Hurtful language knowledge improves sarcasm detection with a neural language model.

摘要

•Hurtful language is consistently found in a corpus of Italian sarcastic language.•Error analysis shows that hostility triggers ironic and sarcastic interpretation.•Negative emotion features improve irony detection with a neural language model.•Hurtful language knowledge improves sarcasm detection with a neural language model.

论文关键词：Affective language,Hurtful language,Irony detection,Sarcasm detection,Linguistic features,AlBERTo

论文评审过程：Received 26 January 2021, Revised 30 November 2021, Accepted 10 December 2021, Available online 3 January 2022, Version of Record 17 January 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.116398