---
title: "Enhancing Bias Detection in Political News Using Pragmatic Presupposition (2020)"
description: "The paper tests whether hidden background assumptions in news text can signal political bias. 96.49% accuracy. 14 references, cited by 4 papers here."
authors: ["Lalitha Kameswari", "Dama Sravani", "Radhika Mamidi"]
year: 2020
venue: "Proceedings of the Eighth International Workshop on Natural Language Processing for Social Media"
source: "https://aclanthology.org/2020.socialnlp-1.1/"
doi: "10.18653/v1/2020.socialnlp-1.1"
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/kameswari-2020-enhancing-bias-detection-political-news/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/kameswari-2020-enhancing-bias-detection-political-news/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2020](https://md.research.elamur.ai/free/en/papers/2020/page/7/) / Enhancing Bias Detection in Political News Using Pragmatic Presupposition

# Enhancing Bias Detection in Political News Using Pragmatic Presupposition

**Authors:** Lalitha Kameswari, Dama Sravani and [Radhika Mamidi](https://md.research.elamur.ai/free/en/authors/radhika-mamidi/) · **Year:** 2020 · **Venue:** [Proceedings of the Eighth International Workshop on Natural Language Processing for Social Media](https://md.research.elamur.ai/free/en/venues/socialnlp/) ([International Workshop on Natural Language Processing for Social Media 2020](https://md.research.elamur.ai/free/en/venues/socialnlp/2020/)) · **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/2020.socialnlp-1.1/) · [PDF](https://aclanthology.org/2020.socialnlp-1.1.pdf) · [DOI 10.18653/v1/2020.socialnlp-1.1](https://doi.org/10.18653/v1/2020.socialnlp-1.1) · [OpenAlex](https://openalex.org/W3045592540)

**Citations:** 3 citations (OpenAlex, 2026-10-14) · [4](https://md.research.elamur.ai/free/en/papers/kameswari-2020-enhancing-bias-detection-political-news/cited-by/) in our library, of the 141,840 papers whose bibliographies we read

## Summary
The paper tests whether hidden background assumptions in news text can signal political bias. The authors focus on pragmatic presupposition: content the writer treats as given but the reader cannot be expected to know. They start from a Telugu political news corpus by Gangula et al. (2019), filter it, and recast its labels as a binary biased/unbiased decision. They also write guidelines for spotting such presuppositions sentence by sentence. Each sentence is marked 0 or 1, and the article's presupposition value is the average of its sentence marks.

Biased articles score clearly higher on this measure than neutral ones. Adding the single presupposition number to bag-of-features vectors lifts every one of six standard classifiers far more than adding the headline does. The best setup surpasses the earlier attention-based model.

The work matters for two reasons:
- It gives computational support to a claim from discourse analysis that had mostly been argued from small hand-picked samples.
- It is a resource for Telugu, a low-resource language.

One caveat: the presupposition values used in the experiments come from human annotation, not from an automatic tagger.

## Key points
- **Corpus:** 1329 Telugu articles from Gangula et al. (2019) were filtered to 1050 (850 biased, 200 unbiased). The original party-specific positive/negative labels were collapsed into bias = 1 / 0. The annotated corpus is publicly released.
- **Guideline cues for marking a sentence as 1:** unresolvable coreference, deixis with no antecedent in the article, Telugu verbal suffixes pointing to unexplained events (-ina, -ani, -tuna, -unTE), verbal nouns of change or continuation of state, and rhetorical questions that hint at non-common-knowledge actions.
- **Presupposition scores:** the mean was 0.46 for biased articles versus 0.15 for unbiased ones. No unbiased article exceeded 0.6, while biased ones reached 1.0.
- **Setup:** chi-square selection reduced features to 10,000. Six scikit-learn classifiers were run with default settings: Bernoulli NB, Multinomial NB, Logistic Regression, RBF SVM, Random Forest and MLP. Evaluation used 10-fold cross-validation over four input variants (article, plus headline, plus presupposition, plus both).
- **Results:** the best was MLP on headline + article + presupposition, with 96.49% accuracy and F1 95.68. This is 6.95% above the attention network of Gangula et al. (2019). Non-linear models (MLP, SVM) led overall.

## Interesting ideas
- The paper scores how much a text assumes rather than states — the share of sentences leaning on unexplained background — and treats this as a measurable signal of a writer's stance.
- Deixis is checked recursively: follow the reference back through the article, and count it as presupposed only if no earlier mention grounds it in context.
- Morphology serves as a presupposition trigger: specific participial and verbal-noun forms signal events the writer takes for granted.
- One cheap, linguistically motivated scalar added to a large sparse vector beat a richer extra signal (the headline) and a neural baseline.
- The annotation unit is a binary sentence decision averaged into a document score, which keeps the guidelines simple and agreement easy to check.

## For Mova

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

- **I1** (frames): prag could compute a per-text "assumed-but-unsupported" ratio from detected presupposition triggers and expose it as an explainable stance or slant indicator.
- **I2** (world): when the world of the text holds no grounded antecedent for an entity, the entity could be flagged as imported background knowledge instead of being silently created.
- **I3** (frames): English analogues visible in the UD tree, such as participial modifiers, nominalized events and "again"/"continue" verbs, could tag frames as presupposed rather than asserted.
- **I4** (learn): when evaluating a pragmatic feature, an ablation that contrasts it with an equally cheap non-linguistic addition is a useful control showing the gain is real.
- **I5** (learn): silver or LLM-teacher labeling of presupposition could use the same sentence-level 0/1 scheme, so that disagreement between witnesses is easy to measure.

*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

- [International Institute of Information Technology, Hyderabad](https://md.research.elamur.ai/free/en/universities/international-institute-of-information-technology-hyderabad/) — University, Hyderabad, India · 24 works · 1,719 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 9,086 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 8,889 works · [🤝 Social Sciences](https://md.research.elamur.ai/free/en/shelves/social-sciences/) · 744 works · [🤝 Sociology and Political Science](https://md.research.elamur.ai/free/en/shelves/social-sciences/sociology-and-political-science/) · 221 works

🧭 Topics (1): [😉 Implicature](https://md.research.elamur.ai/free/en/topics/implicature/) · 72 works

## Cite

Cite as: Lalitha Kameswari, Dama Sravani and Radhika Mamidi. 2020. Enhancing Bias Detection in Political News Using Pragmatic Presupposition. Proceedings of the Eighth International Workshop on Natural Language Processing for Social Media.

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

```bibtex
@inproceedings{kameswari2020enhancing,
  title = {{Enhancing Bias Detection in Political News Using Pragmatic Presupposition}},
  author = {Lalitha Kameswari, Dama Sravani and Radhika Mamidi},
  booktitle = {Proceedings of the Eighth International Workshop on Natural Language Processing for Social Media},
  year = {2020},
  doi = {10.18653/v1/2020.socialnlp-1.1},
  url = {https://doi.org/10.18653/v1/2020.socialnlp-1.1}
}
```

- [← previous Arabic Opinion Mining Using a Hybrid Recommender System Approach](https://md.research.elamur.ai/free/en/papers/harrag-2020-arabic-opinion-mining-using-hybrid/)
- [next → Estimating POS Annotation Consistency of Different Treebanks in a Language](https://md.research.elamur.ai/free/en/papers/aggarwal-2020-estimating-pos-annotation-consistency-different/)
- [↑ back to papers of 2020](https://md.research.elamur.ai/free/en/papers/2020/page/7/)

## References

14 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Enrique Alcarza. 1999. Stylistics in the framework of pragmatics. In ATAS DEL XXI CONGRESO INTER-NACIONAL DE AEDAN, pages 35–54. [find](https://scholar.google.com/scholar?q=Enrique+Alcarza%2E+1999%2E+Stylistics+in+the+framework+of+pragmatics%2E+In+ATAS+DEL+XXI+CONGRESO+INTER%2DNACIONAL+DE+AEDAN%2C+pages+35%E2%80%9354%2E)
2. Mesfin Awoke Bekalu. 2006. Presupposition in news discourse. Discourse & Society, 17(2):147–172. [↗](https://doi.org/10.1177/0957926506060248)
3. Norman Fairclough and Ruth Wodak. 1997. Critical discourse analysis. Discourse studies: A multidisciplinary introduction, 2:258–284. [find](https://scholar.google.com/scholar?q=Norman+Fairclough+and+Ruth+Wodak%2E+1997%2E+Critical+discourse+analysis%2E+Discourse+studies%3A+A+multidisciplinary+introduction%2C+2%3A258%E2%80%93284%2E)
4. Rama Rohit Reddy Gangula, Suma Reddy Duggen-pudi, and Radhika Mamidi. 2019. Detecting political bias in news articles using headline attention. In Proceedings of the 2019 ACL Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP, pages 77–84. [find](https://scholar.google.com/scholar?q=Rama+Rohit+Reddy+Gangula%2C+Suma+Reddy+Duggen%2Dpudi%2C+and+Radhika+Mamidi%2E+2019%2E+Detecting+political+bias+in+news+articles+using+headline+attention%2E+In+Proceedings+of+the+2019+ACL+Workshop+BlackboxNLP%3A+Ana)
5. Lalitha Kameswari and Radhika Mamidi. 2018. Political discourse analysis: A case study of 2014 andhra pradesh state assembly election of interpersonal speech choices. In Proceedings of the 32nd Pacific Asia Conference on Language, Information and Computation. [find](https://scholar.google.com/scholar?q=Lalitha+Kameswari+and+Radhika+Mamidi%2E+2018%2E+Political+discourse+analysis%3A+A+case+study+of+2014+andhra+pradesh+state+assembly+election+of+interpersonal+speech+choices%2E+In+Proceedings+of+the+32nd+Pacifi)
6. Lauri Karttunen and Stanley Peters. 1979. Conventional lmplicature. In Presupposition, pages 1–56. Brill. [find](https://scholar.google.com/scholar?q=Lauri+Karttunen+and+Stanley+Peters%2E+1979%2E+Conventional+lmplicature%2E+In+Presupposition%2C+pages+1%E2%80%9356%2E+Brill%2E)
7. Sandeep Sricharan Mukku, Nurendra Choudhary, and Radhika Mamidi. 2016. Enhanced sentiment classification of telugu text using ml techniques. SAAIP@ IJCAI, 2016:29–34. [find](https://scholar.google.com/scholar?q=Sandeep+Sricharan+Mukku%2C+Nurendra+Choudhary%2C+and+Radhika+Mamidi%2E+2016%2E+Enhanced+sentiment+classification+of+telugu+text+using+ml+techniques%2E+SAAIP%40+IJCAI%2C+2016%3A29%E2%80%9334%2E)
8. Jeppe Nørregaard, Benjamin D Horne, and Sibel Adalı. 2019. Nela-gt-2018: A large multi-labelled news dataset for the study of misinformation in news articles. In Proceedings of the International AAAI Conference on Web and Social Media, volume 13, pages 630–638. [↗](https://doi.org/10.48550/arxiv.1904.01546)
9. Fabian Pedregosa, Gaël Varoquaux, Alexandre Gramfort, Vincent Michel, Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, et al. 2011. Scikit-learn: Machine learning in python. Journal of machine learning research, 12(Oct):2825–2830. [↗](https://doi.org/10.48550/arxiv.1201.0490)
10. Mandy Simons. 2013. On the conversational basis of some presuppositions. In Perspectives on linguistic pragmatics, pages 329–348. Springer. [↗](https://doi.org/10.1007/978-3-319-01014-4_13)
11. Robert Stalnaker. 1974. Pragmatic presupposition//semantics and philosophy. New York, pages 197–214. [find](https://scholar.google.com/scholar?q=Robert+Stalnaker%2E+1974%2E+Pragmatic+presupposition%2F%2Fsemantics+and+philosophy%2E+New+York%2C+pages+197%E2%80%93214%2E)
12. Teun A Van Dijk. 1990. Discourse & society: a new journal for a new research focus. Discourse & Society, 1(1):5–16. [↗](https://doi.org/10.1177/0957926590001001001)
13. Teun A Van Dijk. 1995. Opinions and ideologies in editorials. In 4th International Symposium of Critical Discourse Analysis, Language, Social Life and Critical Thought, Athens, pages 14–16. [find](https://scholar.google.com/scholar?q=Teun+A+Van+Dijk%2E+1995%2E+Opinions+and+ideologies+in+editorials%2E+In+4th+International+Symposium+of+Critical+Discourse+Analysis%2C+Language%2C+Social+Life+and+Critical+Thought%2C+Athens%2C+pages+14%E2%80%9316%2E)
14. Ying-fang Wang. 2010. Analysis of presupposition and its function in advertisement. Canadian Social Science, 3(4):55–60. [↗](https://doi.org/10.3968/j.css.1923669720070304.012)

*Updated 24 October 2026*