---
title: "Rhetorical Move Detection in English Abstracts: Multi-label — paper by Dayrell et al. (2012)"
description: "To handle this, the authors introduce a system called MAZEA, whose name stands for Multi-label Argumentative… 26 references, cited by 12 papers here."
authors: ["Carmen Dayrell", "Arnaldo Candido Jr.", "Gabriel Lima", "Danilo Machado Jr.", "Ann Copestake", "Valéria Feltrim", "Stella Tagnin", "Sandra Aluisio"]
year: 2012
venue: "Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12)"
source: "https://aclanthology.org/L12-1428/"
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/dayrell-2012-rhetorical-move-detection-english-abstracts/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/dayrell-2012-rhetorical-move-detection-english-abstracts/)

[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/) / Rhetorical Move Detection in English Abstracts: Multi-label Sentence Classifiers and their Annotated Corpora

# Rhetorical Move Detection in English Abstracts: Multi-label Sentence Classifiers and their Annotated Corpora

**Authors:** Carmen Dayrell, Arnaldo Candido Jr., Gabriel Lima, Danilo Machado Jr., [Ann Copestake](https://md.research.elamur.ai/free/en/authors/ann-copestake/), Valéria Feltrim, Stella Tagnin and [Sandra Aluisio](https://md.research.elamur.ai/free/en/authors/sandra-aluisio/) · **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/L12-1428/) · [PDF](https://aclanthology.org/L12-1428.pdf) · [OpenAlex](https://openalex.org/W2250845598)

**Citations:** 43 citations (OpenAlex, 2026-10-14) · [12](https://md.research.elamur.ai/free/en/papers/dayrell-2012-rhetorical-move-detection-english-abstracts/cited-by/) in our library, of the 141,840 papers whose bibliographies we read

## Summary
Most earlier systems that tag rhetorical moves in scientific abstracts give each sentence exactly one label. Real sentences often combine functions, for example stating a method and a result together. To handle this, the authors introduce a system called MAZEA, whose name stands for Multi-label Argumentative Zoning for English Abstracts, and it allows a single sentence to receive more than one of six moves: background, gap, purpose, method, result and conclusion.

They built two manually annotated corpora, much larger than those used in earlier work:
- **PE:** 645 abstracts from physical sciences and engineering.
- **LH:** 690 abstracts from life and health sciences.

Each corpus was split in half. One half was used to mine formulaic expressions that are statistically tied to moves. The other half trained multi-label learners from Mulan on top of Weka base classifiers. The features combine these lexical cues with a subset of features from the earlier AZEA system.

Classifier Chain with SMO performed best and clearly beat the random and majority baselines. Its agreement with human annotators came fairly close to the agreement humans reached with each other.

The corpora and a web demo are publicly released as a benchmark for multi-label move detection.

## Key points
- **Corpora:** PE has 645 abstracts (144,683 tokens) and LH has 690 abstracts (50,248 tokens). Both are annotated with six moves and publicly released.
- **How common multi-label sentences are:** 16.5% of LH sentences (1,082 of 6,544) and 11.3% of PE sentences (445 of 3,933).
- **Features (52 numeric in total):**
  - 6 come from formulaic expressions found by the Jbow system.
  - 46 are binary features obtained by expanding 6 AZEA features: position, length, agent expression, modal verb, tense of the first finite verb, and voice.
- **Learners and results:** Classifier Chain + SMO slightly outperformed Rakel + J48. Example-based accuracy on LH was about 0.69, against baselines of 0.1666 (random) and 0.337 (always predicting "method"). PE scores were somewhat lower, in line with weaker human agreement on that corpus.
- **Kappa:**

  | Corpus | Human vs. human | Humans + Chain + SMO |
  |---|---|---|
  | LH | 0.652 | 0.567 |
  | PE | 0.535 | 0.409 |

  Open problems are label order within a sentence and segmenting sentences into move units.

## Interesting ideas
- The authors mined cue phrases on one half of the corpus and trained the classifier on the other half, so lexicon induction and model fitting never see the same texts and the scores are not inflated.
- The system was scored by adding it as an extra annotator in a kappa calculation, which compares it directly against the human-agreement ceiling rather than only against baselines.
- A sentence is allowed several discourse functions instead of being forced into its most salient one, which better matches the fact that moves are functional rather than formal units.
- Automatic clause- or phrase-level segmentation from a parser was dropped after a pilot, because its cut points disagreed with where humans wanted to split. Unit boundaries should be checked against people before a whole annotation scheme is built on them.
- Recurrent formulaic word combinations, weighted by their statistical link to each label, serve as explainable features instead of a bag of words.

## For Mova

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

- **I1** (learn): Self-induced repair rules or cue lexicons should be mined on one silver split and scored on a disjoint split, so the gains are not circular.
- **I2** (learn): Speech-act or role labels from prag could be judged by how much inter-annotator kappa drops when Mova joins the human or LLM-teacher annotators, which shows how far it is from the ceiling.
- **I3** (frames): Speech-act and discourse-function tagging in prag should permit several acts per sentence, for example an assertion plus an evaluation, instead of a single winning label.
- **I4** (tree): Before deriving discourse or frame units from UD clause boundaries, a small pilot should test whether those cuts match human segmentation.
- **I5** (frames): A small table of label-linked formulae could act as transparent cues for detecting gaps, purposes or conclusions in expository texts, and every fired cue can be shown as the explanation.

*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 Cambridge](https://md.research.elamur.ai/free/en/universities/university-of-cambridge/) — University, Cambridge, United Kingdom · 155 works · 115,562 citations
- [Universidade de São Paulo](https://md.research.elamur.ai/free/en/universities/universidade-de-sao-paulo/) — University, São Paulo, Brazil · 3 works · 505 citations

Also on the byline: [Universidade Estadual de Maringá](https://ror.org/04bqqa360).

📚 Shelves (5): [🖥️ 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 · [🎭 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 · [📚 Literature and Literary Theory](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/literature-and-literary-theory/) · 340 works

🧭 Topics (1): [English](https://md.research.elamur.ai/free/en/topics/english/) · 1,373 works

## Cite

Cite as: Carmen Dayrell, Arnaldo Candido Jr., Gabriel Lima, Danilo Machado Jr., Ann Copestake, Valéria Feltrim, Stella Tagnin and Sandra Aluisio. 2012. Rhetorical Move Detection in English Abstracts: Multi-label Sentence Classifiers and their Annotated Corpora. 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{dayrell2012rhetorical,
  title = {{Rhetorical Move Detection in English Abstracts: Multi-label Sentence Classifiers and their Annotated Corpora}},
  author = {Carmen Dayrell, Arnaldo Candido Jr., Gabriel Lima, Danilo Machado Jr., Ann Copestake, Valéria Feltrim, Stella Tagnin and Sandra Aluisio},
  booktitle = {Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12)},
  year = {2012},
  doi = {10.63317/3ecbmfsrdejp},
  url = {https://doi.org/10.63317/3ecbmfsrdejp}
}
```

- [← previous Machine Translation of Labeled Discourse Connectives](https://md.research.elamur.ai/free/en/papers/meyer-2012-machine-translation-labeled-discourse-connectives/)
- [next → Dbnary: Wiktionary as a LMF based Multilingual RDF network](https://md.research.elamur.ai/free/en/papers/serasset-2012-dbnary-wiktionary-as-lmf-based/)
- [↑ back to papers of 2012](https://md.research.elamur.ai/free/en/papers/2012/)

## References

26 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Anthony, L., Lashkia, G. (2003). Mover: A machine learning tool to assist in the reading and writing of technical papers. IEEE Transactions on Professional Communication, 46(3), pp.185--193. [↗](https://doi.org/10.1109/tpc.2003.816789)
2. Carletta, J. (1996). Assessing Agreement on Classification Tasks: The Kappa Statistic. Computational Linguistics, vol. 22, n. 2, pp. 249--254. [↗](https://doi.org/10.48550/arxiv.cmp-lg/9602004) [our card](https://md.research.elamur.ai/free/en/papers/carletta-1996-assessing-agreement-classification-tasks-kappa/)
3. Cortes, Corinna, Vapnik, Vladimir N. (1995). Support-Vector Networks, Machine Learning, 20. [find](https://scholar.google.com/scholar?q=Cortes%2C+Corinna%2C+Vapnik%2C+Vladimir+N%2E+%281995%29%2E+Support%2DVector+Networks%2C+Machine+Learning%2C+20%2E)
4. Feltrim, V. D.; Teufel, S.; Nunes, M. G. V. and Aluísio, S.M. (2006). Argumentative zoning applied to critiquing novices' scientific abstracts. Computing Attitude and Affect in Text: Theory and Applications. [↗](https://doi.org/10.1007/1-4020-4102-0_18)
5. The Information Retrieval Series, 2006, Volume 20, pp. 233--246. [find](https://scholar.google.com/scholar?q=The+Information+Retrieval+Series%2C+2006%2C+Volume+20%2C+pp%2E+233%2D%2D246%2E)
6. Genovês Jr., L.; Feltrim, V.D.; Dayrell, C. and Aluísio, S. (2007). Automatically detecting schematic structure components of English abstracts. In Proceedings of the RANLP´2007, Workshop on Natural Language Processing for Educational Resources. Borovets, Bulgaria, pp. 23--29. [find](https://scholar.google.com/scholar?q=Genov%C3%AAs+Jr%2E%2C+L%2E%3B+Feltrim%2C+V%2ED%2E%3B+Dayrell%2C+C%2E+and+Alu%C3%ADsio%2C+S%2E+%282007%29%2E+Automatically+detecting+schematic+structure+components+of+English+abstracts%2E+In+Proceedings+of+the+RANLP%C2%B42007%2C+Workshop+on+Natural+L)
7. Hirohata, K.; Okazaki, N.; Ananiadou, S. and Ishizuka, M. (2008). Identifying sections in scientific abstracts using conditional random fields. In Proceedings of the 3rd International Joint Conference on Natural Language Processing (IJCNLP 2008). Asian Federation of Natural Language Processing, pp. 381--388. [find](https://scholar.google.com/scholar?q=Hirohata%2C+K%2E%3B+Okazaki%2C+N%2E%3B+Ananiadou%2C+S%2E+and+Ishizuka%2C+M%2E+%282008%29%2E+Identifying+sections+in+scientific+abstracts+using+conditional+random+fields%2E+In+Proceedings+of+the+3rd+International+Joint+Conference)
8. Hyland, K. (2004). Disciplinary Discourses. Michigan: The University of Michigan Press. [find](https://scholar.google.com/scholar?q=Hyland%2C+K%2E+%282004%29%2E+Disciplinary+Discourses%2E+Michigan%3A+The+University+of+Michigan+Press%2E)
9. Ito, T.; Simbo, M.; Yamasaki, T. and Matsumoto, Y. (2004). Semi-supervised sentence classification for medline documents. In IPSJ SIG Technical Report, v. 2004-ICS-138, pp. 141--146. [find](https://scholar.google.com/scholar?q=Ito%2C+T%2E%3B+Simbo%2C+M%2E%3B+Yamasaki%2C+T%2E+and+Matsumoto%2C+Y%2E+%282004%29%2E+Semi%2Dsupervised+sentence+classification+for+medline+documents%2E+In+IPSJ+SIG+Technical+Report%2C+v%2E+2004%2DICS%2D138%2C+pp%2E+141%2D%2D146%2E)
10. Lafferty, J.; McCallum, A. and Pereira, F (2001). Conditional random Fields: Probabilistic models for segmenting and labelling sequence data. In Proceedings of the International Conference on Machine learning (ICML), Williams College, Williamstown, MA, USA, Morgan Kaufmann, pp. 282--289. [find](https://scholar.google.com/scholar?q=Lafferty%2C+J%2E%3B+McCallum%2C+A%2E+and+Pereira%2C+F+%282001%29%2E+Conditional+random+Fields%3A+Probabilistic+models+for+segmenting+and+labelling+sequence+data%2E+In+Proceedings+of+the+International+Conference+on+Machine+)
11. Landis, J. R. , Koch, G. G (1977). The measurement of observer agreement for categorical data. Biometrics, v. 33, pp. 159--174. [↗](https://doi.org/10.2307/2529310)
12. Lin, J.; Karakos, D.; Demner-fushman, D. and Khudanpur, S. (2006). Generative content models for structural analysis of medical abstracts. In Proceedings of the HLT/NAACL 2006 Workshop on Biomedical Natural Language Processing (BioNLP’06), Association for Computational Linguistics (ACL), pp. 65--72. [find](https://scholar.google.com/scholar?q=Lin%2C+J%2E%3B+Karakos%2C+D%2E%3B+Demner%2Dfushman%2C+D%2E+and+Khudanpur%2C+S%2E+%282006%29%2E+Generative+content+models+for+structural+analysis+of+medical+abstracts%2E+In+Proceedings+of+the+HLT%2FNAACL+2006+Workshop+on+Biomedical+N)
13. Machado Jr., D. (2009). Extração Automática de Expressões Indicativas para a Classificação de Textos Científicos. In Proceedings of The 7th Brazilian Symposium in Information and Human Language Technology, I TILIC, 2009 (In Portuguese). [find](https://scholar.google.com/scholar?q=Machado+Jr%2E%2C+D%2E+%282009%29%2E+Extra%C3%A7%C3%A3o+Autom%C3%A1tica+de+Express%C3%B5es+Indicativas+para+a+Classifica%C3%A7%C3%A3o+de+Textos+Cient%C3%ADficos%2E+In+Proceedings+of+The+7th+Brazilian+Symposium+in+Information+and+Human+Language+Techno)
14. McKnight, L., Arinivasan, P. (2003). Categorization of sentence types in medical abstracts. In AMIA 2003 Symposium Proceedings, American Medical Informatics Association, pp. 440--444. [find](https://scholar.google.com/scholar?q=McKnight%2C+L%2E%2C+Arinivasan%2C+P%2E+%282003%29%2E+Categorization+of+sentence+types+in+medical+abstracts%2E+In+AMIA+2003+Symposium+Proceedings%2C+American+Medical+Informatics+Association%2C+pp%2E+440%2D%2D444%2E)
15. Pendar, N. , Cotos, E. (2008). Automatic Identification of Discourse Moves in Scientific Article Introductions. In Proceedings of The Third Workshop on Innovative Use of NLP for Building Educational Applications, Association for Computational Linguistics (ACL), pp. 62-70. [find](https://scholar.google.com/scholar?q=Pendar%2C+N%2E+%2C+Cotos%2C+E%2E+%282008%29%2E+Automatic+Identification+of+Discourse+Moves+in+Scientific+Article+Introductions%2E+In+Proceedings+of+The+Third+Workshop+on+Innovative+Use+of+NLP+for+Building+Educational+A)
16. Read, J.; Pfahringer, B.; Holmes, G. and Frank, E., (2009). Classifier Chains for Multi-label Classification. In Proceedings of the Machine Learning and Knowledge Discovery in Databases, Proceedings, Part II, Lecture Notes in Computer Science, Vol. 5782, Subseries: Lecture Notes in Artificial Intelligence, Buntine, W.; Grobelnik, M.; Mladenic, D.; Shawe-Taylor, J. (Eds.), pp. 254--269. [find](https://scholar.google.com/scholar?q=Read%2C+J%2E%3B+Pfahringer%2C+B%2E%3B+Holmes%2C+G%2E+and+Frank%2C+E%2E%2C+%282009%29%2E+Classifier+Chains+for+Multi%2Dlabel+Classification%2E+In+Proceedings+of+the+Machine+Learning+and+Knowledge+Discovery+in+Databases%2C+Proceedings%2C+)
17. Ruch, P.; Boyer, C.; Chichester, C.; Tbahriti, I.; Geissbuhler, A.; Fabry, P.; Gobeill, J.; Pillet, V.; Rebholz-Schuhmann, D.; Lovis, C. and Veuthey, A. L. (2007). Using argumentation to extract key sentences from biomedical abstracts. International Journal of Medical Informatics, v. 76, pp. 195--200. [↗](https://doi.org/10.1016/j.ijmedinf.2006.05.002)
18. Shimbo, M.; Yamasaki, T. and Matsumoto, Y. (2003). Using sectioning information for text retrieval: a case study with the medline abstracts. In Proceedings of the 2nd International Workshop on Active Mining (AM’03), Maebashi, Japan; 2003 , pp. 32--41. [find](https://scholar.google.com/scholar?q=Shimbo%2C+M%2E%3B+Yamasaki%2C+T%2E+and+Matsumoto%2C+Y%2E+%282003%29%2E+Using+sectioning+information+for+text+retrieval%3A+a+case+study+with+the+medline+abstracts%2E+In+Proceedings+of+the+2nd+International+Workshop+on+Active+)
19. Swales, J. (2004). Research Genres: Exploration and applications. Cambridge University Press, Cambridge. [find](https://scholar.google.com/scholar?q=Swales%2C+J%2E+%282004%29%2E+Research+Genres%3A+Exploration+and+applications%2E+Cambridge+University+Press%2C+Cambridge%2E)
20. Swales J. M. and Feak, C. B. (2009). Abstracts and the Writing of Abstracts. Michigan: University of Michigan Press. [find](https://scholar.google.com/scholar?q=Swales+J%2E+M%2E+and+Feak%2C+C%2E+B%2E+%282009%29%2E+Abstracts+and+the+Writing+of+Abstracts%2E+Michigan%3A+University+of+Michigan+Press%2E)
21. Teufel, S., Moens, M. (2002). Summarising scientific articles experiments with relevance and rhetorical status. Computational Linguistics 28 (4), pp. 409--446. [find](https://scholar.google.com/scholar?q=Teufel%2C+S%2E%2C+Moens%2C+M%2E+%282002%29%2E+Summarising+scientific+articles+experiments+with+relevance+and+rhetorical+status%2E+Computational+Linguistics+28+%284%29%2C+pp%2E+409%2D%2D446%2E)
22. Tsoumakas, G., Vlahavas, I. (2007). Random k-Labelsets: An Ensemble Method for Multilabel Classification, In Proceedings of the 18th European Conference on Machine Learning (ECML 2007), Lecture Notes in Artificial Intelligence 4701, Joost N. Kok, Jacek Koronacki, Ramon Lopez de Mantaras, Stan Matwin, Dunja Mladenic (Eds), pp. 406--417. [find](https://scholar.google.com/scholar?q=Tsoumakas%2C+G%2E%2C+Vlahavas%2C+I%2E+%282007%29%2E+Random+k%2DLabelsets%3A+An+Ensemble+Method+for+Multilabel+Classification%2C+In+Proceedings+of+the+18th+European+Conference+on+Machine+Learning+%28ECML+2007%29%2C+Lecture+Notes+)
23. Tsoumakas, G.; Katakis, I. and Vlahavas, I. (2010). Mining Multi-label Data. Data Mining and Knowledge Discovery Handbook, O. Maimon, L. Rokach (Ed.), Springer, 2nd edition,. [↗](https://doi.org/10.1007/978-0-387-09823-4_34)
24. Witten, I. H., Frank, E. (2005). Data mining: practical machine learning tools and techniques. 2nd ed. San Francisco: Morgan Kaufmann. [find](https://scholar.google.com/scholar?q=Witten%2C+I%2E+H%2E%2C+Frank%2C+E%2E+%282005%29%2E+Data+mining%3A+practical+machine+learning+tools+and+techniques%2E+2nd+ed%2E+San+Francisco%3A+Morgan+Kaufmann%2E)
25. Wu, J.; Chang, Y.; Liou, H. and Chang, J. S. (2006). Computational analysis of move structures in academic abstracts. In Proceedings of the COLING/ACL 2006 Interactive Presentation Sessions, pp. 41--44. [find](https://scholar.google.com/scholar?q=Wu%2C+J%2E%3B+Chang%2C+Y%2E%3B+Liou%2C+H%2E+and+Chang%2C+J%2E+S%2E+%282006%29%2E+Computational+analysis+of+move+structures+in+academic+abstracts%2E+In+Proceedings+of+the+COLING%2FACL+2006+Interactive+Presentation+Sessions%2C+pp%2E+41%2D%2D4)
26. Yamamoto, Y., Takagi, T. (2005). A sentence classification system for multi-document summarization in the biomedical domain. In Proceedings of the 21st International Conference on Data Engineering Workshops (ICDEW'05), IEEE Computer Society Washington, DC, USA, pp. 90--95. [find](https://scholar.google.com/scholar?q=Yamamoto%2C+Y%2E%2C+Takagi%2C+T%2E+%282005%29%2E+A+sentence+classification+system+for+multi%2Ddocument+summarization+in+the+biomedical+domain%2E+In+Proceedings+of+the+21st+International+Conference+on+Data+Engineering+Wor)

*Updated 24 October 2026*