---
title: "Computational Sarcasm — Bhattacharyya 2017"
description: "This is a tutorial abstract from EMNLP 2017, not a research paper with experiments. Cited by 1 paper here."
authors: ["Pushpak Bhattacharyya", "Aditya Joshi"]
year: 2017
venue: "Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing: Tutorial Abstracts"
source: "https://aclanthology.org/D17-3002/"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/bhattacharyya-2017-computational-sarcasm/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/bhattacharyya-2017-computational-sarcasm/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2017](https://md.research.elamur.ai/free/en/papers/2017/page/4/) / Computational Sarcasm

# Computational Sarcasm

**Authors:** [Pushpak Bhattacharyya](https://md.research.elamur.ai/free/en/authors/pushpak-bhattacharyya/) and [Aditya Joshi](https://md.research.elamur.ai/free/en/authors/aditya-joshi/) · **Year:** 2017 · **Venue:** [Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing: Tutorial Abstracts](https://md.research.elamur.ai/free/en/venues/emnlp/) ([EMNLP 2017](https://md.research.elamur.ai/free/en/venues/emnlp/2017/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/D17-3002/) · [PDF](https://aclanthology.org/D17-3002.pdf) · [OpenAlex](https://openalex.org/W3088036339)

**Citations:** 0 citations (OpenAlex, 2026-10-14) · 1 in our library, of the 141,840 papers whose bibliographies we read

## Summary
This is a tutorial abstract from EMNLP 2017, not a research paper with experiments. It outlines a three-hour course on computational sarcasm, meaning automatic methods for handling sarcastic text. The motivation is that sarcasm undermines sentiment analysis, because the literal polarity of a sarcastic utterance often contradicts what the speaker means. The tutorial begins with linguistic theory and treats incongruity as the central signal behind sarcasm and related figurative language. It then focuses on sarcasm detection, framed as a binary decision of whether a text is sarcastic. It covers:
- how datasets are built and annotated, and the pitfalls involved;
- rule-based, feature-driven statistical, topic-model and deep learning detectors;
- the use of context beyond the target text, from the conversation or from the author's history;
- newer tasks: sarcasm generation, separating sarcasm from irony, and identifying the target of sarcasm.

No new results are reported. The value of the tutorial is as a structured map of the field. It draws on the authors' survey (arXiv:1602.03426) and on the authors' own prior detectors.

## Key points
- Planned length is 180 minutes in 11 parts, with a 95-minute block and an 85-minute block, ending in a 5-minute live demonstration of the authors' detector.
- Detection methods are grouped into four families: rule-based (one specific sarcasm cue applied as a rule), statistical classifiers with engineered features, a topic-model approach, and deep learning.
- The dataset section names two annotation problems: supervision from hashtags and the effect of annotators who are not native speakers.
- Two kinds of context are covered: conversational context (10 minutes) and the author's historical context (5 minutes).
- The "beyond detection" topics are sarcasm vs. irony classification, sarcasm generation and sarcasm target identification. The source survey is Joshi, Bhattacharyya & Carman (2016), arXiv:1602.03426.

## Interesting ideas
- Incongruity is the unifying cue for sarcasm: a clash between parts of an utterance, or between the utterance and the situation, which also underlies other figurative language.
- In rule-based detection, each rule encodes one specific piece of evidence for sarcasm, so every decision traces back to a named cue.
- Labels taken from hashtags are cheap but noisy, and non-native annotators can systematically misjudge sarcasm, so the way labels are produced should itself be examined.
- Sarcasm often cannot be decided from the sentence alone, so evidence comes from earlier conversation turns or the author's past statements.
- Finding the target of sarcasm is treated as its own task, separate from deciding whether sarcasm is present at all.

## For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

- **I1** (frames): crate prag could flag irony when the sentiment of a frame contradicts the state of the world model, and the flag would be explainable because it points to the two clashing facts.
- **I2** (frames): irony detection in prag could be built as a set of named, inspectable cue rules instead of an opaque score, which fits the requirement that every step be explainable.
- **I3** (learn): when silver irony labels come from heuristics or LLM teachers, Mova should check for annotator-type bias and keep negative controls to catch systematically wrong labels.
- **I4** (world): the world of the text could store each speaker's earlier stances, letting the irony check compare a new utterance against what that character previously said or believes.
- **I5** (answer): in QA on fables and tales, knowing who is being mocked lets a reader answer questions about a character's real attitude instead of the literal words.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [Indian Institute of Technology Bombay](https://md.research.elamur.ai/free/en/universities/indian-institute-of-technology-bombay/) — University, Mumbai, India · 22 works · 2,018 citations
- [Commonwealth Scientific and Industrial Research Organisation](https://md.research.elamur.ai/free/en/government/commonwealth-scientific-and-industrial-research-organisation/) — Government body, Canberra, Australia · 5 works · 1,634 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 9,085 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 8,888 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 9,296 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 9,206 works

🧭 Topics (1): [😊 Sentiment & emotion](https://md.research.elamur.ai/free/en/topics/sentiment-emotion/) · 609 works

## Cite

Cite as: Pushpak Bhattacharyya and Aditya Joshi. 2017. Computational Sarcasm. Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing: Tutorial Abstracts.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{bhattacharyya2017computational,
  title = {{Computational Sarcasm}},
  author = {Pushpak Bhattacharyya and Aditya Joshi},
  booktitle = {Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing: Tutorial Abstracts},
  year = {2017},
  url = {https://aclanthology.org/D17-3002/}
}
```

- [← previous The translation technique and the translation of verbal forms in Psalm 84 of the Masoretic Text as employed by the Peshitta](https://md.research.elamur.ai/free/en/papers/moretsi-2017-translation-technique-translation-verbal-forms/)
- [next → Crossroads between contrastive linguistics, translation studies and machine translation: TC3-II](https://md.research.elamur.ai/free/en/papers/czulo-2017-crossroads-between-contrastive-linguistics/)
- [↑ back to papers of 2017](https://md.research.elamur.ai/free/en/papers/2017/page/4/)

*Updated 24 October 2026*