Production of Demonstratives in Dutch, English and Portuguese Dialogues

What it is about

The paper reports an elicitation study of indexical determiners (this/that/these/those, Dutch deze/die, Portuguese este/aquele) used with a pointing gesture. Participants read a scripted furniture-shop dialogue, in which a seller introduces five first-mention referents, and they chose a definite article, a proximal or a distal demonstrative for each. The study has 184 native speakers: 91 English, 51 Dutch and 42 Portuguese, giving 920 demonstratives and determiners. Two hypotheses are tested: accessibility (H1: Dutch uses proximals for hard-to-access referents, English and Portuguese the opposite) and distance (H2: proximal for near, distal for far). The distance hypothesis fits better in all three languages. For English, the distal is chosen for the far referent RE2 (81%), the proximal for the near-middle referent RE4 (69%) and the distal for RE5 (67%). For the middle referent RE3, English and Dutch speakers mostly chose the definite article rather than a demonstrative, which the authors read as the accessibility of an object previously mentioned as "the desk next to it".

Ideas

• I1. A rule in the generator for English indexical determiners: a first-mention referent that is marked as pointed at takes this/these if the speaker-distance feature is near, that/those if far, and the definite article if the referent was introduced in the immediately preceding clause. Keep the three choices as an explicit enum, so each output says which rule fired.
• I2. Distance as a feature of the world: a position for each entity in world (near, middle, far from the speaker), derived from a text description ("on the left, far from the door"). The paper's table gives a check set for it.
• I3. Use the paper's exception as a rule: a referent given by "the X next to it" is accessible, so the definite article wins over the demonstrative for English. Measure how often the rule gives the definite article on the RE3 type of item.
• I4. Negative control: shuffle the distance labels across referents. The rule must score worse than with the true labels on the English table; otherwise the distance feature has no effect.

For Mova

The chapters are 12. Pronouns, determiners and numerals (the demonstrative and definite determiners, the near-far contrast) and 36. Discourse relations and coherence or 35. Reference and coreference (the accessibility of a referent). The code is coref/src/mention.rs (mention type: definite, demonstrative, pronoun) and world/src/story.rs (entities and their locations). The measure: rule agreement with the English percentages in the paper (RE2 distal, RE4 proximal, RE5 distal), run twice for identical output, plus the shuffled-distance negative control. The study data are evaluation-only, so Mova's own English test items should be made from its texts.

Cite

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

@inproceedings{luz2011production,
  title = {{Production of Demonstratives in Dutch, English and Portuguese Dialogues}},
  author = {Luz, Saturnino and Sluis, Ielka van der},
  booktitle = {Proceedings of the 13th European Workshop on Natural Language Generation},
  year = {2011},
  url = {https://aclanthology.org/W11-2825/}
}

References

15 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. S. Anderson and E. Keenan. 1985. Deixis. In Language Typology and Syntactic Fieldwork Vol. III. Cambridge: Cambridge University Press. find
2. E. André, T. Rist, S. Mulken, and M. Van Klesen. 2000. The automated design of believable dialogues for animated presentation teams. In Embodied Conversational Agents. MIT Press. ↗
3. A. Bangerter. 2004. Using pointing and describing to achieve joint focus of attention in dialogue. Psychological Science, 15:415–419. ↗
4. K. Bühler. 1934. Sprachtheorie: Die Darstellungsfunktion der Sprache. Fischer, Jena. find
5. D. Byron and L. Stoia. 2005. An analysis of proximity markers in collaborative dialog. In Proceedings of CLS-05. find
6. M. Cavalcante. 2002. O demonstrativo e seus usos. Perspectiva, 20(1). find
7. H. Clark and A. Bangerter. 2004. Changing ideas about reference. In I. Noveck and D. Sperber, editors, Experimental Pragmatics, pages 25–49. Palgrave Macmillan, New York. ↗
8. H. Clark. 1996. Using Language. Cambridge University Press. find
9. J.K. Gundel, N. Hedberg, and R. Zacharski. 1993. Cognitive status and the form of referring expressions in discourse. Language, 69(2):274–307. ↗
10. K. Jungbluth. 2005. Os pronomes demonstrativos do português brasileiro na fala e na escrita. Cadernos de Linguagem e Sociedade, 7:83–105. find
11. P. Piwek and A. Cremers. 1996. Dutch and English demonstratives: A comparison. Language Sciences, 18(3-4):835–851. ↗
12. P. Piwek, R. Beun, and A. Cremers. 2008. ‘Proximal’ and ‘distal’ in language and cognition: Evidence from deictic demonstratives in Dutch. Journal of Pragmatics, 40:694–718. ↗
13. I. Van der Sluis and S. Luz. 2011a. A cross-linguistic study on the production of multimodal referring expressions in dialogue. In Proceedings of the 13th European Workshop on Natural Language Generation, Nancy, France. find
14. I. Van der Sluis and S. Luz. 2011b. Issues in translating and producing Japanese referring expressions for dialogues. Linguistic Issues in Language Technology, 5(1):1–46. I. Van der Sluis, S. Luz W. Breitfuß, M. Ishizuka, and H. Prendinger. to appear. Cross-cultural assessment of automatically generated multimodal referring expressions in a virtual world. International Journal of Human-Computer Studies. ↗
15. S. Williams, P. Piwek, and R. Power. 2007. Generating monologue and dialogue to present personalised medical information to patients. In Proc. of ENLG-07. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/luz-2011-production-demonstratives-dutch-english-portuguese/