«Are you Afraid of Ghosts?» A Proposal for Busting Predicate Ellipsis in Universal Dependencies

Summary

The paper deals with predicate ellipsis — a missing verb in a clause — and how to annotate it in Universal Dependencies treebanks. The basic UD layer has no empty nodes, so ellipsis is only worked around: one remaining element is promoted to the head position, or the orphan relation links the remnants, and the gap itself is never marked explicitly. The authors propose two complementary workflows: a theoretical one for recognising ellipsis (locating the gap, the remnants and the antecedent) and a practical one for restoring the missing predicate as an empty node in the Enhanced Dependencies layer. The approach is tested on the Italian-Old treebank, which contains Dante's Divine Comedy, and difficult cases are discussed, such as where to place the restored node and what to do when no antecedent is available in context. The authors stress the shortage of gold data with explicit ellipsis for both linguistic analysis and machine learning, and present the work as a first step towards language-independent UD guidelines for ellipsis.

How Mova uses it

• en/seeds/errors/gapping-orphan.md: rule en.errors.orphan-head — in gapping ("Marie went to Paris and Miriam to Prague") an orphan must hang on the promoted remnant (a conjunct, root, advcl or parataxis); otherwise the gapping was analysed wrongly.
• en/seeds/errors/vp-ellipsis-aux-head.md: rule en.errors.aux-before-verb — in VP ellipsis ("Mary will too") the stranded auxiliary becomes the clause head; an aux placed after its verb signals that the elided clause was wrongly attached as an auxiliary of the first verb.
• The paper is cited for the basic-layer convention (promotion or orphan, empty nodes only in Enhanced UD), which these rules check in the expert system (engine en/src/expert.rs).

Effectiveness in Mova

Not measured separately. The rules it supports are validated on gold treebanks: on UD English EWT 2.18, 29 orphan relations with 0 violations (GUM: 128 / 6); aux after a verbal head — 2 cases (inversion), GUM — 1. Role: source for one pair of annotation-consistency rules.

👨‍🔬💥

Topics: Universal Dependencies · Treebanks

Cite

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

@inproceedings{corbetta2025are,
  title = {{«Are you Afraid of Ghosts?» A Proposal for Busting Predicate Ellipsis in Universal Dependencies}},
  author = {Corbetta, Claudia and Iurescia, Federica and Passarotti, Marco Carlo},
  booktitle = {Proceedings of the 23rd International Workshop on Treebanks and Linguistic Theories (TLT, SyntaxFest 2025)},
  year = {2025},
  url = {https://aclanthology.org/2025.tlt-1.6/}
}

References

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

1. Guillaume Bonfante, Bruno Guillaume, and Guy Perrier. 2018. Application of Graph Rewriting to Natural Language Processing. John Wiley & Sons. find
2. Damir Ćavar, Ludovic Mompelat, and Muhammad Abdo. 2024a. The typology of ellipsis: a corpus for linguistic analysis and machine learning applications. In Proceedings of the 6th Workshop on Research in Computational Linguistic Typology and Multilingual NLP, pages 46–54. find
3. Damir Ćavar, Zoran Tiganj, Ludovic Veta Mompelat, and Billy Dickson. 2024b. Computing ellipsis constructions: Comparing classical nlp and llm approaches. In Proceedings of the Society for Computation in Linguistics 2024, pages 217–226. find
4. Claudia Corbetta, Marco Passarotti, Flavio Massimiliano Cecchini, and Giovanni Moretti. 2023. Highway to hell. towards a universal dependencies treebank for dante alighieri’s comedy. find
5. Marie-Catherine De Marneffe, Christopher D Manning, Joakim Nivre, and Daniel Zeman. 2021. Universal dependencies. Computational linguistics, 47(2):255– 308. find our card
6. Kira Droganova, Filip Ginter, Jenna Kanerva, and Daniel Zeman. 2018. Mind the gap: Data enrichment in dependency parsing of elliptical constructions. In Proceedings of the Second Workshop on Universal Dependencies (UDW 2018), pages 47–54. find our card
7. Gaël Guibon, Marine Courtin, Kim Gerdes, and Bruno Guillaume. 2020. When collaborative treebank curation meets graph grammars. In LREC 2020-12th Language Resources and Evaluation Conference. find
8. Jan Hajič, Eva Hajičová, Marie Mikulová, and Jiří Mírovskỳ. 2017. Handbook on linguistic annotation, chapter prague dependency treebank. find
9. Jorge Hankamer and Ivan Sag. 1976. Deep and surface anaphora. Linguistic inquiry, 7(3):391–428. find
10. Philip Hofmeister. 2007. Memory retrieval effects on filler-gap procession. In Proceedings of the Annual Meeting of the Cognitive Science Society, volume 29. find
11. John Robert Ross. 1969. Guess who? In Proceedings from the annual meeting of the chicago linguistic society, volume 5, pages 252–286. Chicago Linguistic Society. find
12. Richard Hudson. 2010. An introduction to word grammar. Cambridge University Press. find
13. Maria Simi, Simonetta Montemagni, et al. 2018. Bootstrapping enhanced universal dependencies for italian. In CEUR WORKSHOP PROCEEDINGS, volume 2253. CEUR-WS. find
14. Christopher Kennedy. 2003. Ellipsis and syntactic representation. In The interfaces: Deriving and interpreting omitted structures, pages 29–53. John Benjamins Publishing Company. find
15. Jeroen Van Craenenbroeck and Tanja Temmerman. 2018. The Oxford handbook of ellipsis. Oxford University Press. find
16. Thomas S. Kuhn. 1962. The Structure of Scientific Revolutions. University of Chicago Press. find
17. A Mitch Marcus, Beatrice Santorini, and Mary Ann Marcinkiewicz. 1993. Building a large annotated corpus of english: The penn treebank. Computational linguistics, 19(2):313–330. find our card

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/corbetta-2025-guillemotleftare-you-afraid-ghosts-guillemotright/