Swedish Prepositions are not Pure Function Words
Summary
The paper examines whether Swedish prepositions should be classed as "pure function words". In UD, this term refers to a special class of words that are almost never allowed to have dependents. UD justifies this status for prepositions by their similarity to case endings. The author looks at Swedish sentences in which prepositions (på, i) play different roles. These are an adverbial, a verb particle, a prepositional object, a preposition without its noun, a preposition before an infinitive or a subordinate clause, and a preposition with an adverbial modifier. For each case, the author compares the current UD guidelines with linguistic intuition. The author also compares how prepositions are actually annotated in the Scandinavian, English and German treebanks. It turns out that treebanks of closely related languages annotate these words in very different ways. Prepositions with adverbial modifiers occur in all of them. The author concludes that Swedish prepositions are not pure function words. Whether a word can have dependents depends on the particular word rather than on its part of speech. The work matters for consistent annotation and for refining the part-of-speech categories in UD.
Key points
- The UD guidelines for prepositions are as follows:
- obl + case for adverbials and prepositional objects;
- compound:prt for verb particles;
- mark for a preposition before a VP or a subordinate clause (a decision from the discussion of issue #257 on the UD forum).
- In the construction "mitt på näsan", UD attaches the modifier to the noun rather than to the preposition. The author considers this wrong, because the phrase becomes ungrammatical without the preposition. Moreover, subordinating conjunctions (SCONJ) are allowed to have modifiers.
- 7 treebanks are compared: Danish, English, German, Norwegian Bokmål, Norwegian Nynorsk, Swedish and Swedish LinES. The share of the mark relation in them ranges from 0.000 (LinES) to 0.115 (Bokmål).
- The Norwegian treebanks tag the particles bort, hjem, inn, opp, ut as ADP, while the Swedish ones tag them as ADV. The author attributes this to differing annotator guidelines rather than to differences between the languages. All treebanks contain prepositions with advmod, for example "omedelbart efter lunch" in LinES.
- The author proposes describing UD parts of speech as sets of properties with gradience, drawing on the model of Aarts (2007). The author also proposes merging ADP and SCONJ and, where needed, marking the difference between them with a feature such as AdpType.
Cite
@inproceedings{ahrenberg2017swedish,
title = {{Swedish Prepositions are not Pure Function Words}},
author = {Ahrenberg, Lars},
booktitle = {Proceedings of the NoDaLiDa 2017 Workshop on Universal Dependencies (UDW 2017)},
year = {2017},
url = {https://aclanthology.org/W17-0402/}
}References
- Bas Aarts. 2007. Syntactic Gradience: The Nature of Grammatical Indeterminacy. Oxford University Press, Oxford. find
- Dwight L. Bolinger. 1971. The Phrasal Verb in English. Harvard University Press, Cambridge, MA. find
- Joseph E. Emonds. 1976. A Transformational Approach to English Syntax. Academic Press, New York. find
- Joseph E. Emonds. 1985. A Unified Theory of Syntactic Categories. Foris, Dordrecht. ↗
- Marie-Catherine De Marneffe, Timothy Dozat, Natalia Silveira, Katri Haverinen, Filip Ginter, Joakim Nivre, and Christopher D. Manning. 2014. Universal stanford dependencies: a cross-linguistic typology. In Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC’14), Reykjavik, Iceland, may. find our card
- Slav Petrov, Dipanjan Das, and Ryan McDonald. 2012. A universal part-of-speech tagset. In Proceedings of the Eigth International Conference on Language Resources and Evaluation (LREC 2016), Istanbul, Turkey, may. find
- Joakim Nivre, Marie-Catherine de Marneffe, Filip Ginter, Yoav Goldberg, Jan Hajič, Christopher D. Manning, Ryan McDonald, Slav Petrov, Sampo Pyysalo, Natalia Silveira, Reut Tsarfaty, and Daniel Zeman. 2016. Universal dependencies v1: A multilingual treebank collection. In Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC 2016). find our card
- Ulf Teleman, Erik Andersson, and Staffan Hellberg. 2010. Svenska Akademins Grammatik. Norstedts, Stockholm. find
- Joakim Nivre, Željko Agić, Lars Ahrenberg, Maria Jesus Aranzabe, Masayuki Asahara, Aitziber Atutxa, Miguel Ballesteros, John Bauer, Kepa Bengoetxea, Riyaz Ahmad Bhat, Eckhard Bick, Cristina Bosco, Gosse Bouma, Sam Bowman, Marie Candito, Gülşen Cebirolu Eryiit, Giuseppe G. A. Celano, Fabricio Chalub, Jinho Choi, Çar Çöltekin, Miriam Connor, Elizabeth Davidson, Marie-Catherine de Marneffe, Valeria de Paiva, Arantza Diaz de Ilarraza, Kaja Dobrovoljc, Timothy Dozat, Kira Droganova, Puneet Dwivedi, Marhaba Eli, Tomaž Erjavec, Richárd Farkas, Jennifer Foster, Cláudia Freitas, Katarı́na Gajdošová, Daniel Galbraith, Marcos Garcia, Filip Ginter, Iakes Goenaga, Koldo Gojenola, Memduh Gökrmak, Yoav Goldberg, Xavier Gómez Guinovart, Berta Gonzáles Saavedra, Matias Grioni, Normunds Grūzītis, Bruno Guillaume, Nizar Habash, Jan Hajič, Linh Hà M, Dag Haug, Barbora Hladká, Petter Hohle, Radu Ion, Elena Irimia, Anders Johannsen, Fredrik Jørgensen, Hüner Kaşkara, Hiroshi UD. 2017a. Core dependents in ud v2. http://universaldependencies.org/v2/core-dependents.html. ↗
- UD. 2017b. Morphology: General principles. http://universaldependencies.org/u/overview/morphology.html. ↗
- UD. 2017c. Syntax: General principles. http://universaldependencies.org/u/overview/syntax.html. ↗
This page in Markdown → https://md.research.elamur.ai/free/papers/ahrenberg-2017-swedish-prepositions-are-not-pure/