Elamur Research⌕
☰
home / papers / 2013 / Using the Semantic-Syntactic Interface for Reliable Arabic Modality Annotation

Using the Semantic-Syntactic Interface for Reliable Arabic Modality Annotation

Paper · 2013 · Rania Al-Sabbagh, Jana Diesner and Roxana Girju

Summary

Arabic had no corpus annotated for modality. Even the Penn Arabic Treebank carries no such labels. The authors build the first one. It uses a random sample of political tweets from the YADAC corpus, posted in Egypt during the first half-year of the 2011 revolution and written in both Egyptian Arabic (EA) and MSA.

Their central contribution is a two-part definition of a modality trigger:

  • the expression must carry a meaning from Palmer's typology (epistemic, evidential, deontic, dynamic and their subtypes), and
  • it must take a clause or a verbal phrase as its complement.

Semantics keeps ordinary clause-taking words out, and syntax excludes words that only look modal in meaning. Two annotators labeled candidate triggers for meaning, scope type and scope span. Fully matching decisions produced 1,343 valid triggers and 1,034 rejected candidates, and scope spans reached a macro kappa of 0.929. The work shows that anchoring a semantic category in a syntactic frame yields guidelines that others can replicate. It also yields concrete features for later automatic detection of triggers and their scopes.

Key points

  • Corpus: 2,377 candidate triggers on which both annotators fully agreed (meaning, scope type, span): 1,343 accepted as modal, 1,034 rejected. Individually, the annotators accepted 1,746 and 1,619.
  • Scope span agreement is high (macro κ = 0.929), which the authors credit to short, telegraphic tweets. Disagreements stemmed from interjections, adjuncts, subordinate conjunctions and scopes recoverable only through implicature.
  • Scope type tracks modality class: clauses are the usual complement for propositional modality (epistemic, evidential), whereas verbal phrases are typical for event modality (deontic, dynamic). Implicit scopes make up just 2% of the total, and roughly 82% of scope heads appear right beside their trigger.
  • Ambiguity is common. 37% of valid modal lemmas are lexically ambiguous (modal or not, e.g. qAl, $Af), and 5% are semantically ambiguous (several modal readings, e.g. Erf, mmkn as epistemic vs abilitive).
  • Dialect split of valid triggers, by AML labels: 77.3% EA-only, 15.6% either variety depending on context, about 7% MSA-only.

Interesting ideas

  • I1. A category is admitted only when meaning and syntactic frame both agree, so each criterion cancels the other's false positives and the guideline becomes checkable.
  • I2. The observed link between modality class and scope type (propositional → clause, event → verbal phrase) can serve as a consistency check on labels.
  • I3. Rejected candidates that annotators agreed on are kept as an explicit negative set rather than discarded.
  • I4. Implicit scopes are sorted by how they are recovered: a deictic pointing to a prior clause, an elided repeated complement, or a missing light verb before a noun phrase.
  • I5. Agreement is counted only on exact matches across all layers at once (meaning, scope type, span), not per layer.

For Mova

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

  • I1 (frames): a modal frame could be opened only when the head has a modal sense and also governs a ccomp/xcomp in the UD tree, with the failed condition reported as the explanation.
  • I2 (frames): if an "epistemic" reading attaches to a bare VP, or a "deontic" reading to a full clause, the frame builder could flag the case for review or prefer the other sense of an ambiguous lemma like "must".
  • I3 (learn): store agreed non-modal look-alikes (e.g. "say", "see" in non-modal uses) as negative controls when evaluating a modality detector or self-induced rules.
  • I4 (frames): coreference could link "we won't allow it" to the antecedent clause, filling the frame's scope slot with a proposition rather than a pronoun.
  • I5 (learn): silver data from witness parsers could require joint agreement on trigger, frame and span before an example is accepted, which gives stricter but cleaner training material.

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

🎓University of Illinois Urbana-ChampaignUrbana · United States209 works · 71,394 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 works

🧭 Topics 1Research subjects and languages the work is about.Semitic languages · 3,066 works

Cite

Cite as Rania Al-Sabbagh, Jana Diesner and Roxana Girju. 2013. Using the Semantic-Syntactic Interface for Reliable Arabic Modality Annotation. Proceedings of the Sixth International Joint Conference on Natural Language Processing.

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

@inproceedings{alsabbagh2013using,
  title = {{Using the Semantic-Syntactic Interface for Reliable Arabic Modality Annotation}},
  author = {Rania Al-Sabbagh, Jana Diesner and Roxana Girju},
  booktitle = {Proceedings of the Sixth International Joint Conference on Natural Language Processing},
  year = {2013},
  url = {https://aclanthology.org/I13-1047/}
}

References

20 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Muhammad Abdul-Mageed and Mona Diab. 2012. AWATIF: A Multi-Genre Corpus for Modern Standard Arabic Subjectivity and Sentiment Analysis. The 8th International Conference on Language Resources and Evaluation (LREC'12), Istanbul, Turkey, 21-27 May 2012.
  2. Rania Al-Sabbagh and Roxana Girju. 2012. YADAC: Yet another Dialectal Arabic Corpus. The 8th International Conference on Language Resources and Evaluation (LREC'12), Istanbul, Turkey, 21-27 May 2012.
  3. Kathrin Baker, Michael Bloodgood, Bonnie Dorr, Nathaniel W. Filardo, Lori Levin, and Christine Piatko. 2010. A Modality Lexicon and its Use in Automatic Tagging. The 7th International Conference on Language Resources and Evaluation (LREC'10), Valletta, Malta, 19-21 May 2010. find
  4. Virginia Brennan. 1993. Root and Epistemic Modal Auxiliary Verbs, PhD Thesis, University of Massachusetts, Amherst. find
  5. Kristen E. Brustad. 2000. The Syntax of Spoken Arabic: A Comparative Study of Moroccan, Egyptian, Syrian and Kuwaiti Dialects. Georgetown University Press, Washington DC, USA. find
  6. John Butler. 2003. A Minimalist Treatment of Modality. Lingua, vol. 113 (10): 967-996. ↗
  7. Joan L. Bybee, R. D. Perkins and W. Pagliuca. 1994. The Evolution of Grammar: Tense, Aspect and Modality in the Languages of the World. University of Chicago Press, Chicago, USA. find
  8. Mona Diab, Lori Levin, Teruko Mitamura, Owen Rambow, Vinodkumar Prabhakaran and Wei Wei Guo. 2009. Commited Belief Annotation and Tagging. Proceedings of the Third Linguistic Annotation Workshop (ACL-IJCNLP '09), August 2009, Suntec, Singapore. find
  9. Iris Hendrickx, Amália Mendes, and Silvia Mencarelli. 2010. Modality in Text: A Proposal for Corpus Annotation. The 7th International Conference on Language Resources and Evaluation (LREC'10), Valletta, Malta, 19-21 May 2010 find
  10. Ray Jackendoff. 1972. Semantic Interpretation in Generative Grammar. Cambridge, MA: The MIT Press. find
  11. Suguru, Matsuyoshi, Megumi Eguchi, Chitose Sao, Koji Murakami, Kentaro Inui, and Yuji Matsumoto. 2010. Annotating Event Mentions in Text with Modality, Focus, and Source Information. The 7th International Conference on Language Resources and Evaluation (LREC'10), Valletta, Malta, 19-21 May 2010 find
  12. T. F. Mitchell and S. A. Al-Hassan. 1994. Modality, Mood and Aspect in Spoken Arabic with Special Reference to Egypt and the Levant. London and NY: Kegan Paul International. find
  13. Ola Moshref. 2012. Corpus Study of Tense, Aspect, and Modality in Diaglossic Speech in Cairene Arabic. PhD Thesis. University of Illinois at Urbana-Champaign. find
  14. Frank R. Palmer. 1989. Mood and Modality. Cambridge University Press, Cambridge, UK. ↗
  15. Frank R. Palmer. 2001. Mood and Modality. 2 nd Edition. Cambridge University Press, Cambridge, UK. ↗
  16. Vinodkumar Prabhakaran. 2012. Detecting Power Relations from Written Dialog. Proceedings of the 2012 Student Research Workshop, Jeju, Republic of Korea. find
  17. Vinodkumar Prabhakaran, Michael Bloodgood, Mona Diab, Bonnie Dorr, Lori Levin, Christine D. Piatko, Owen Rambow and Benjamin Van Durme. 2012. Statistical Modality Tagging from Rule-Based Annotations and Crowdsourcing. Proceedings of the ACL-2012 Workshop on Extra-Propositional Aspects of Meaning in Computational Linguistics (ExProM-2012), Jeju, Republic of Korea. find
  18. Roser Saurí, March Verhagen and James Pustejovsky. 2006. Annotating and Recognizing Event Modality in Text. The 19th International FLAIRS Conference, pages: 335-339, Florida, USA, May 2006 find
  19. György Szarvas, Veronika Vincze, Richárd Farkas, and János Csirik. 2008. The BioScope Corpus: Annotation for Negation, Uncertainty and their Scope in Biomedical Texts. BioNLP 2008: Current Trends in Biomedical Natural Language Processing, pages: 38-45, Columbus, Ohio, USA, June 2008
  20. Janyce Wiebe, Theresa Wilson, and Claire Cardie. 2005. Annotating Expressions of Opinions and Emotions in Language. Language Resources and Evaluations, 39:165-210 ↗

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/sabbagh-2013-semantic-syntactic-interface-reliable-arabic/