Abstract / Description of output
Sarcasm detection models are often built based on self-annotated tagged data. However, fillers (e.g., um and hmm), deliberate use of which may indicate sarcasm, do not get enough attention in these models. We analyze five fillers in different categories of untagged tweets. We also present participant ratings of sarcasm, offensive language, language formality, and basic emotions in tweets with and without um and hmm. Our evidence, albeit weak, points to the importance of linguistic features such as these fillers in determining sarcastic meaning.
Original language | English |
---|---|
Title of host publication | Proceedings of the 26th Workshop on the Semantics and Pragmatics of Dialogue - Poster Abstracts |
Editors | Eleni Gregoromichelaki, Julian Hough, John D. Kelleher |
Publisher | SEMDIAL |
Pages | 1-3 |
Number of pages | 3 |
Publication status | Published - 24 Aug 2022 |
Event | The 26th Workshop on the Semantics and Pragmatics of Dialogue - Technological University Dublin, Dublin, Ireland Duration: 22 Aug 2022 → 24 Aug 2022 Conference number: 26 https://semdial2022.github.io/ |
Publication series
Name | Proceedings of the Workshop Series on the Semantics and Pragmatics of Dialogue |
---|---|
Publisher | SEMDIAL |
ISSN (Electronic) | 2308-2275 |
Workshop
Workshop | The 26th Workshop on the Semantics and Pragmatics of Dialogue |
---|---|
Abbreviated title | SemDial 2022 |
Country/Territory | Ireland |
City | Dublin |
Period | 22/08/22 → 24/08/22 |
Internet address |