---
title: "Clause-based Discourse Segmentation of Arabic Texts (2012)"
description: "The paper tackles the first step of discourse parsing for Arabic: splitting text into Elementary Discourse… 19 references, cited by 4 papers here."
authors: ["Iskandar Keskes", "Farah Benamara", "Lamia Hadrich Belguith"]
year: 2012
venue: "Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12)"
source: "https://aclanthology.org/L12-1559/"
licence: "CC BY-NC 4.0 (ELRA, LREC proceedings)"
licence_url: "https://creativecommons.org/licenses/by-nc/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/keskes-2012-clause-based-discourse-segmentation-arabic/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/keskes-2012-clause-based-discourse-segmentation-arabic/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2012](https://md.research.elamur.ai/free/en/papers/2012/page/2/) / Clause-based Discourse Segmentation of Arabic Texts

# Clause-based Discourse Segmentation of Arabic Texts

**Authors:** [Iskandar Keskes](https://md.research.elamur.ai/free/en/authors/iskandar-keskes/), [Farah Benamara](https://md.research.elamur.ai/free/en/authors/farah-benamara/) and [Lamia Hadrich Belguith](https://md.research.elamur.ai/free/en/authors/lamia-hadrich-belguith/) · **Year:** 2012 · **Venue:** [Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12)](https://md.research.elamur.ai/free/en/venues/lrec/) ([LREC 2012](https://md.research.elamur.ai/free/en/venues/lrec/2012/)) · **Publisher:** [European Language Resources Association (ELRA)](https://md.research.elamur.ai/free/en/publishers/european-language-resources-association-elra/) · **Licence:** [CC BY-NC 4.0 (ELRA, LREC proceedings)](https://creativecommons.org/licenses/by-nc/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/L12-1559/) · [PDF](https://aclanthology.org/L12-1559.pdf) · [OpenAlex](https://openalex.org/W2251811147)

**Citations:** 16 citations (OpenAlex, 2026-10-14) · [4](https://md.research.elamur.ai/free/en/papers/keskes-2012-clause-based-discourse-segmentation-arabic/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
The paper tackles the first step of discourse parsing for Arabic: splitting text into Elementary Discourse Units (EDUs). Here those units are sentences, clauses, and some prepositional and noun phrases. Arabic makes this hard for several reasons. Clitics attach to words, there are no capital letters, most text lacks diacritics (so words are highly ambiguous), and punctuation is used irregularly. The authors studied a corpus of news articles and elementary school textbooks, segmented by hand through consensus of three linguists. From this they built a rule-based segmenter with three principles: punctuation only (P1), lexical cues only (P2), and both combined (P3). Rules are written as NooJ local grammars (finite-state transducers) on top of NooJ Arabic dictionaries and morpho-syntactic analysis. Combining cues works best, reaching an F-measure of 85.5% on textbooks and 68% on news. Punctuation alone performs poorly. Error analysis shows that most failures come from lexical and morphological ambiguity, which suggests that semantic disambiguation of discourse cues is the next step.

## Key points
- **Corpus:**
  - 150 news articles: 737 paragraphs, 40,532 words.
  - 250 elementary school textbook (EST) texts from 4th–8th grade: 1,095 paragraphs, 29,473 words.
  - 80% of news and 60% of textbooks were used to derive patterns; the rest was held out for testing.
- **Punctuation classes:**
  - Strong indicators always end a segment: `!`, `?`, `:`, `;`.
  - Weak indicators need decision rules: full stop, comma, quotes, parentheses, brackets, braces, underscores.
  - Example rule: a dot inside a named entity or abbreviation is not a boundary.
- **Results (F-measure, P1 / P2 / P3):**
  - Textbooks: 45% / 66% / 85.5%.
  - News: 21% / 53.5% / 68%.
- **Implementation:** the system runs on the NooJ platform (Silberztein), with rules encoded as local grammars.
- **Error sources:**
  - Abbreviations at the end of a segment.
  - An adverb mis-tagged as a verb.
  - A named entity wrongly split into a clitic plus a verb.
  - Cases where both punctuation and connectors are absent.

## Interesting ideas
- Each punctuation mark is sorted into "always a boundary" or "boundary only under context", and contextual rules are written only for the second group; this keeps the rule set small and easy to audit.
- Connectors are grouped by whether corpus evidence shows they tend to open a segment, close one, or neither. This turns a vague cue list into a typed lexicon.
- The method is evaluated as a cumulative ladder (punctuation, then cues, then both), which shows exactly how much each information source contributes.
- Results are reported separately for simple, well-punctuated school texts and for complex news, which exposes how genre affects ambiguity.
- Errors are traced back to the upstream tagger or analyzer, not blamed on the segmentation rules themselves.

## For Mova

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

- **I1** (tokens): Sentence splitting in crate en could adopt an explicit strong/weak mark table, with a named rule (abbreviation, named entity, single non-verb word before a dot) recorded for each non-split.
- **I2** (frames): Clause-level EDU boundaries from typed English connectors could define the units that prag turns into separate who-did-what frames, with the connector type kept as an explainable link.
- **I3** (learn): Pipeline evaluations could report the same staged ablations, so that each added component must justify itself against a weaker baseline acting as a control.
- **I4** (answer): School-course and fable QA scores should be tracked apart from harder genres, so that easy-text success is not mistaken for general robustness.
- **I5** (learn): Self-induced repair rules could be generated from such attributions, with each segmentation error routed to the stage (tagging, parsing) that actually caused it.

*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

- [Institut de Recherche en Informatique de Toulouse](https://md.research.elamur.ai/free/en/labs/institut-de-recherche-en-informatique-de-toulouse/) — Research institute, Toulouse, France · 36 works · 2,193 citations
- [Faculty of Economics and Management of Sfax](https://md.research.elamur.ai/free/en/universities/faculty-of-economics-and-management-of-sfax/) — University, Sfax, Tunisia · 3 works · 20 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 18,614 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 17,965 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 17,450 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 17,296 works

🧭 Topics (2): [Semitic languages](https://md.research.elamur.ai/free/en/topics/semitic-languages/) · 3,066 works · [🧵 Discourse analysis](https://md.research.elamur.ai/free/en/topics/discourse-analysis/) · 471 works

## Cite

Cite as: Iskandar Keskes, Farah Benamara and Lamia Hadrich Belguith. 2012. Clause-based Discourse Segmentation of Arabic Texts. Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12).

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

```bibtex
@inproceedings{keskes2012clause,
  title = {{Clause-based Discourse Segmentation of Arabic Texts}},
  author = {Iskandar Keskes, Farah Benamara and Lamia Hadrich Belguith},
  booktitle = {Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12)},
  year = {2012},
  doi = {10.63317/26588rymf6s8},
  url = {https://doi.org/10.63317/26588rymf6s8}
}
```

- [← previous Annotating the Focus of Negation in terms of Questions Under Discussion](https://md.research.elamur.ai/free/en/papers/anand-2012-annotating-focus-negation-terms-questions/)
- [next → Open-Source Boundary-Annotated Corpus for Arabic Speech and Language Processing](https://md.research.elamur.ai/free/en/papers/brierley-2012-open-source-boundary-annotated-corpus/)
- [↑ back to papers of 2012](https://md.research.elamur.ai/free/en/papers/2012/page/2/)

## References

19 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Afantenos, S. D.; Denis, P.; Muller, P. and Danlos, L. (2010). [Learning recursive segments for discourse parsing](https://md.research.elamur.ai/free/en/papers/afantenos-2010-learning-recursive-segments-discourse-parsing/). In Proceedings of the International Conference on Language Resources and Evaluation, LREC 2010 (Valletta, Malta, 2010). [↗](https://doi.org/10.48550/arxiv.1003.5372)
2. Asher, N.; Lascarides, A. (2003). Logics Conversation. Cambridge University Press. of Basha, A. Z. (1912)." O0F ا "c$ ا & 8J !\Dو W0 ا. (Punctuation and its marks in Arabic Language). [find](https://scholar.google.com/scholar?q=Asher%2C+N%2E%3B+Lascarides%2C+A%2E+%282003%29%2E+Logics+Conversation%2E+Cambridge+University+Press%2E+of+Basha%2C+A%2E+Z%2E+%281912%29%2E%22+O0F+%D8%A7+%22c%24+%D8%A7+%26+8J+%21%5CD%D9%88+W0+%D8%A7%2E+%28Punctuation+and+its+marks+in+Arabic+Language%29%2E)
3. Belguith Hadrich, L. (2009). Analyse et résumé automatiques de documents : Problèmes, conception et réalisation, Habilitation Universitaire en Informatique, soutenue le 2 mai 2009, FSEGS, Université de Sfax, Tunisie. [find](https://scholar.google.com/scholar?q=Belguith+Hadrich%2C+L%2E+%282009%29%2E+Analyse+et+r%C3%A9sum%C3%A9+automatiques+de+documents+%3A+Probl%C3%A8mes%2C+conception+et+r%C3%A9alisation%2C+Habilitation+Universitaire+en+Informatique%2C+soutenue+le+2+mai+2009%2C+FSEGS%2C+Universit%C3%A9+d)
4. Belguith Hadrich, L. ; Aloulou, C. and Ben Hamadou A. (2008). «MASPAR : De la segmentation à l'analyse syntaxique de textes arabes», Information Interaction Intelligence I3, CÉPADUÈS-Editions, mai 2008, Vol. 7, n° 2, p. 9-36. [find](https://scholar.google.com/scholar?q=Belguith+Hadrich%2C+L%2E+%3B+Aloulou%2C+C%2E+and+Ben+Hamadou+A%2E+%282008%29%2E+%C2%ABMASPAR+%3A+De+la+segmentation+%C3%A0+l%27analyse+syntaxique+de+textes+arabes%C2%BB%2C+Information+Interaction+Intelligence+I3%2C+C%C3%89PADU%C3%88S%2DEditions%2C+mai+200)
5. Belguith, L.; Baccour, L. and Mourad, G. (2005). Segmentation de textes arabes basée sur l'analyse contextuelle des signes de ponctuations et de certaines particules. 12th Conference on Natural Language Processing (TALN’2005), Dourdan. [find](https://scholar.google.com/scholar?q=Belguith%2C+L%2E%3B+Baccour%2C+L%2E+and+Mourad%2C+G%2E+%282005%29%2E+Segmentation+de+textes+arabes+bas%C3%A9e+sur+l%27analyse+contextuelle+des+signes+de+ponctuations+et+de+certaines+particules%2E+12th+Conference+on+Natural+Langua)
6. Chai, J. Y. and Jin, R. (2004). Discourse structure for context question answering. In HLT-NAACL Workshop on Pragmatics of Question Answering. [find](https://scholar.google.com/scholar?q=Chai%2C+J%2E+Y%2E+and+Jin%2C+R%2E+%282004%29%2E+Discourse+structure+for+context+question+answering%2E+In+HLT%2DNAACL+Workshop+on+Pragmatics+of+Question+Answering%2E)
7. Da Cunha, I.; SanJuan, E. and Torres M. (2010). Discourse segmentation for Spanish based on shallow parsing. Proceedings of the 9th Mexican international conference on Advances in artificial intelligence, (MICAI'10), p. 13-23. Springer-Verlag. [find](https://scholar.google.com/scholar?q=Da+Cunha%2C+I%2E%3B+SanJuan%2C+E%2E+and+Torres+M%2E+%282010%29%2E+Discourse+segmentation+for+Spanish+based+on+shallow+parsing%2E+Proceedings+of+the+9th+Mexican+international+conference+on+Advances+in+artificial+intellige)
8. Debili, F. ; Achour, H. and Souissi, E. (2002). La langue arabe et l’ordinateur, de l’étiquetage grammatical à la voyellation automatique. Correspondances n° 71 juillet-août 2002. [find](https://scholar.google.com/scholar?q=Debili%2C+F%2E+%3B+Achour%2C+H%2E+and+Souissi%2C+E%2E+%282002%29%2E+La+langue+arabe+et+l%E2%80%99ordinateur%2C+de+l%E2%80%99%C3%A9tiquetage+grammatical+%C3%A0+la+voyellation+automatique%2E+Correspondances+n%C2%B0+71+juillet%2Dao%C3%BBt+2002%2E)
9. Halliday, M. A. K. & Hasan, R. (1976). Cohesion in English. London: Longman. [find](https://scholar.google.com/scholar?q=Halliday%2C+M%2E+A%2E+K%2E+%26+Hasan%2C+R%2E+%281976%29%2E+Cohesion+in+English%2E+London%3A+Longman%2E)
10. Harald, L.; Csilla, P.; Maja, B.; Mirco, H. and Henning L. (2006). Discourse segmentation of German written text. In: Tapio Salakoski, Filip Ginter, Sampo Pyysalo, Tapio Pahikkala (eds.): Proceedings of the 5th International Conference on Natural Language Processing (FinTAL 2006). Berlin: Springer, 2006. [find](https://scholar.google.com/scholar?q=Harald%2C+L%2E%3B+Csilla%2C+P%2E%3B+Maja%2C+B%2E%3B+Mirco%2C+H%2E+and+Henning+L%2E+%282006%29%2E+Discourse+segmentation+of+German+written+text%2E+In%3A+Tapio+Salakoski%2C+Filip+Ginter%2C+Sampo+Pyysalo%2C+Tapio+Pahikkala+%28eds%2E%29%3A+Proceedings+)
11. Jirawan, C.; Thana, S.; and Asanee K. (2005). "Element Discourse Unit Segmentation for Thai Discourse Cues and Syntactic Information", The 9th National Computer Science and Engineering Conference, 27-28 October, 2005. [find](https://scholar.google.com/scholar?q=Jirawan%2C+C%2E%3B+Thana%2C+S%2E%3B+and+Asanee+K%2E+%282005%29%2E+%22Element+Discourse+Unit+Segmentation+for+Thai+Discourse+Cues+and+Syntactic+Information%22%2C+The+9th+National+Computer+Science+and+Engineering+Conference%2C+27%2D)
12. Khalifa, I.; Feki, Z. and Farawila, A. (2011). Arabic Discourse Segmentation Based on Rhetorical Methods. International Journal of Electric and Computer Sciences IJECS-IJENS, Vol: 11(1). [find](https://scholar.google.com/scholar?q=Khalifa%2C+I%2E%3B+Feki%2C+Z%2E+and+Farawila%2C+A%2E+%282011%29%2E+Arabic+Discourse+Segmentation+Based+on+Rhetorical+Methods%2E+International+Journal+of+Electric+and+Computer+Sciences+IJECS%2DIJENS%2C+Vol%3A+11%281%29%2E)
13. Maâloul, M. H. ; Ellouze, M. and Belguith Hadrich, L. (2008). Al Lakas s El'eli / & {ا ص Aّ$ ا: Un système de résumé automatique de documents arabes. 9th International Business Information Management Conference, IBIMA’08, Marrakech, Maroc, 4-6 janvier 2008. pp 1260 – 1268. [find](https://scholar.google.com/scholar?q=Ma%C3%A2loul%2C+M%2E+H%2E+%3B+Ellouze%2C+M%2E+and+Belguith+Hadrich%2C+L%2E+%282008%29%2E+Al+Lakas+s+El%27eli+%2F+%26+%7B%D8%A7+%D8%B5+A%D9%91%24+%D8%A7%3A+Un+syst%C3%A8me+de+r%C3%A9sum%C3%A9+automatique+de+documents+arabes%2E+9th+International+Business+Information+Management+)
14. Mann, W.C. and Thompson, S. (1988). [Rhetorical Structure Theory: Toward a Functional Theory of Text Organization](https://md.research.elamur.ai/free/en/papers/mann-1988-rhetorical-structure-theory-functional/). Text 8(3). p. 243-281. [↗](https://doi.org/10.1515/text.1.1988.8.3.243)
15. Marcu, D. (2000). The Theory and Practice of Discourse Parsing and Summarization. The MIT Press. [↗](https://doi.org/10.7551/mitpress/6754.001.0001)
16. Mesfar, S. (2008). Analysis morpho-syntaxique automatic recognition of named entites in Standard Arabic. PHD thesis, University of Franche-Comté, France. [find](https://scholar.google.com/scholar?q=Mesfar%2C+S%2E+%282008%29%2E+Analysis+morpho%2Dsyntaxique+automatic+recognition+of+named+entites+in+Standard+Arabic%2E+PHD+thesis%2C+University+of+Franche%2DComt%C3%A9%2C+France%2E)
17. Seeger, F. and Brian, R. (2007). The utility of parse-derived features for automatic discourse segmentation. In ACL. [find](https://scholar.google.com/scholar?q=Seeger%2C+F%2E+and+Brian%2C+R%2E+%282007%29%2E+The+utility+of+parse%2Dderived+features+for+automatic+discourse+segmentation%2E+In+ACL%2E)
18. Silberztein, M. (1993). Dictionnaires électroniques et analyse automatique des textes : Le système INTEX. Masson–Paris. [find](https://scholar.google.com/scholar?q=Silberztein%2C+M%2E+%281993%29%2E+Dictionnaires+%C3%A9lectroniques+et+analyse+automatique+des+textes+%3A+Le+syst%C3%A8me+INTEX%2E+Masson%E2%80%93Paris%2E)
19. Touir, A.; Mathkour, H. and Al-Sanea, W. (2008). Semantic-Based Segmentation of Arabic Texts. Information Technology Journal. Vol: 7(7). [↗](https://doi.org/10.3923/itj.2008.1009.1015)

*Updated 1 November 2026*