Meaning Representation of English Prepositional Phrase Roles: SNACS Supersenses vs. Tectogrammatical Functors

What it is about

The paper compares two inventories for the meaning of prepositional phrases: SNACS supersenses (52 labels in three branches — circumstance: Locus, Time, Duration, Source, Goal, Direction, Manner, Purpose…; participant: Agent, Theme, Recipient, Beneficiary, Instrument, Ancillary, Cost…; configuration: relations between nominals like QuantityItem, Possessor), each preposition carrying a scene role and a function role, and the functors of the Prague English Dependency Treebank (LOC, TWHEN, THL, DIR1/DIR3, MEANS, ACMP, BEN…). From the guidelines they write a heuristic functor→supersense mapping (DIR1 → Source, DIR3 → Goal, THL → Duration), then run a SNACS classifier on 1,837 PPs in 838 WSJ sentences of PEDT. Agreement with the mapping is about 50% overall, best for temporal (66.5%) and agent/patient-like ones, worst for configurations (36%). Mismatches come partly from classifier errors (money read as Cost, a month as Locus) and partly from real differences: fictive motion («60 miles from Seoul» is a Locus scene with a Source function), with + partner as Agent scene with Ancillary function.

Ideas

I1. Give Mova's prepositional obliques a small set of supersenses chosen by deterministic rules instead of the bare preposition (obls holds with=queen): Source (from, out of), Goal (to/into/onto after a motion verb), Locus (in/on/at/under + place), Time and Duration (for + time noun), Instrument (with + inanimate), Ancillary/companion (with + animate), Beneficiary (for + animate), Topic (about), Purpose (for + event noun), Cost (for + money). The rule and the head's animacy go into How. Then «with whom», «what with», «for whom», «from where» questions map to one slot. Measure frame bench obliques/place F1 and the to-whom/with-whom course questions (world knowledge diagnose).
I2. Keep scene and function apart where they diverge, as SNACS does: «a cave a mile from the village» is where the cave is (Locus), not where something came from; with + animate under a verb of joint action (fight with, play with) is a co-agent, not an instrument. Write these as reader tests, each with the wrong reading as the negative control.
I3. Measure the rule set of I1 against STREUSLE's gold supersenses (tests only, outside the repo, like PropBank), on scene roles, per preposition. It shows which prepositions need more than one rule.

For Mova

22. Phrases: noun, adjective, adverb and prepositional phrases (prepositional phrases), 23-complementation.md, 30-pragmatics-overview.md §5.3. Code: prag/src/frames.rs (obliques), prag/src/vn.rs (VerbNet PP roles: Destination, Source, Instrument already used for place/not-place), world/src/knowledge.rs. Measures: frame bench place/oblique F1, course questions with prepositional wh-phrases, per-preposition accuracy on STREUSLE.

Cite

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

@inproceedings{scivetti2023meaning,
  title = {{Meaning Representation of English Prepositional Phrase Roles: SNACS Supersenses vs. Tectogrammatical Functors}},
  author = {Scivetti, Wesley and Schneider, Nathan},
  booktitle = {Proceedings of the Fourth International Workshop on Designing Meaning Representations},
  year = {2023},
  url = {https://aclanthology.org/2023.dmr-1.7/}
}

References

15 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

Adam Meyers, Ruth Reeves, Catherine Macleod, Rachel Szekely, Veronika Zielinska, Brian Young, and Ralph Grishman. 2004. The Nombank Project: An Interim Report. In Proceedings of the Workshop Frontiers in Corpus Annotation at HLT-NAACL 2004, page 24–31, Boston, Massachusetts, USA. Association for Computational Linguistics. find
Aryaman Arora. 2023. snacs: Models for parsing SNACS datasets. https://github.com/ aryamanarora/snacs . ↗
Collin F. Baker, Charles J. Fillmore, and John B. Lowe. 1998. The Berkeley FrameNet project. In 36th Annual Meeting of the Association for Computational Linguistics and 17th International Conference on Computational Linguistics, Volume 1, pages 86–90, Montreal, Quebec, Canada. Association for Computational Linguistics. find our card
Martha Palmer, Claire Bonial, and Jena D. Hwang. 2017. VerbNet: Capturing English verb behavior, meaning, and usage. In Susan E. F. Chipman, editor, The Oxford Handbook of Cognitive Science, pages 315–336. Oxford University Press. find
Ann Bies, Justin Mott, Colin Warner, and Seth Kulick. 2012. English Web Treebank. LDC2012T13. find
Nathan Schneider, Jena D. Hwang, Vivek Srikumar, Archna Bhatia, Na-Rae Han, Tim O’Gorman, Sarah R. Moeller, Omri Abend, Adi Shalev, Austin Blodgett, and Jakob Prange. 2022. Adposition and Case Supersenses v2.6: Guidelines for English. arXiv:1704.02134 [cs.CL]. ↗ our card
Claire Bonial, Julia Bonn, Kathryn Conger, Jena D. Hwang, and Martha Palmer. 2014. PropBank: Semantics of New Predicate Types. In Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC’14), pages 3013– 3019, Reykjavik, Iceland. European Language Resources Association (ELRA). find our card
Nathan Schneider, Jena D. Hwang, Vivek Srikumar, Meredith Green, Abhijit Suresh, Kathryn Conger, Tim O’Gorman, and Martha Palmer. 2016. A Corpus of Preposition Supersenses. In Proceedings of the 10th Linguistic Annotation Workshop held in conjunction with ACL 2016 (LAW-X 2016), pages 99–109, Berlin, Germany. Association for Computational Linguistics. find our card
Jan Hajič, Eva Hajičová, Jarmila Panevová, Petr Sgall, Ondřej Bojar, Silvie Cinková, Eva Fučíková, Marie Mikulová, Petr Pajas, Jan Popelka, Jiří Semecký, Jana Šindlerová, Jan Štěpánek, Josef Toman, Zdeňka Urešová, and Zdeněk Žabokrtský. 2012. Announcing Prague Czech-English Dependency Treebank 2.0. In Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC’12), pages 3153–3160, Istanbul, Turkey. European Language Resources Association (ELRA). find
Nathan Schneider, Jena D. Hwang, Vivek Srikumar, Jakob Prange, Austin Blodgett, Sarah R. Moeller, Aviram Stern, Adi Bitan, and Omri Abend. 2018. Comprehensive supersense disambiguation of English prepositions and possessives. In Proc. of ACL, pages 185–196, Melbourne, Australia. find
Jena D. Hwang, Archna Bhatia, Na-Rae Han, Tim O’Gorman, Vivek Srikumar, and Nathan Schneider. 2017. Double Trouble: The Problem of Construal in Semantic Annotation of Adpositions. In Proceedings of the 6th Joint Conference on Lexical and Computational Semantics (*SEM 2017), page 178–188, Vancouver, Canada. Association for Computational Linguistics. find
Nathan Schneider, Vivek Srikumar, Jena D. Hwang, and Martha Palmer. 2015. A hierarchy with, of, and for preposition supersenses. In Proc. of The 9th Linguistic Annotation Workshop, pages 112–123, Denver, Colorado, USA. find our card
Jena D. Hwang, Nathan Schneider, and Vivek Srikumar. 2020. Sprucing up Supersenses: Untangling the Semantic Clusters of Accompaniment and Purpose. In Proceedings of the 14th Linguistic Annotation Workshop, page 127–137, Barcelona, Spain. Association for Computational Linguistics. find
Leonard Talmy. 1996. Fictive motion in language and “ception”. In Paul Bloom, Mary A. Peterson, Nadel Lynn, and Merrill F. Garrett, editors, Language and Space, pages 211–276. MIT Press, Cambridge, MA. find
Karin Kipper, Anna Korhonen, Neville Ryant, and Martha Palmer. 2008. A Large-scale Classification of English Verbs. Language Resources and Evaluation, 42(1):21–40. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/scivetti-2023-meaning-representation-english-prepositional-phrase/