The neural machine translation of dislocations

Authors

  • Behnoosh Namdarzadeh University of Tehran, Iran Author
  • Nicolas Ballier Sorbonne University, France Author

DOI:

https://doi.org/10.36505/ExLing-2022/13/0035/000577

Keywords:

dislocation, French-English language pairs, neural machine translation

Abstract

his paper investigates neural machine translation (NMT) outputs for dislocated constructions translated from French into English. Dislocations are often considered “substandard in formal registers” (Lambrecht 1994: 12). In French, multiple copies of the subject are permissible in spoken discourse, whereas equivalent constructions are generally excluded in English translations (De Cat 2007). We analysed 436 translations of French dislocated segments from Voyage au bout de la nuit and spoken-language data from the Corpus de Français Parlé Parisien (CFPP) using DeepL and Google Translate. Beyond prototypical X, c’est dislocations, translation systems continue to produce inaccurate outputs, potentially due to the limited inclusion of spoken-language data in NMT training corpora.

References

Branca-Rosoff, S., Lefeuvre, F. 2016. Le Corpus de Français Parlé Parisien des années 2000: Constitution, outils et analyses. Le cas des interrogatives indirectes. Corpus, 15, 265-284. https://doi.org/10.4000/corpus.3043

Louis-Ferdinand Céline 1932. Voyage au bout de la nuit. Paris: Gallimard.

De Cat, C. 2007. French Dislocation: Interpretation, Syntax, Acquisition. Oxford: Oxford University Press.

Isabelle, P., Cherry, C., Foster, G. 2017. A Challenge Set Approach to Evaluating Machine Translation. In Proceedings of the Association for Computational Linguistics, 2486-2496. https://doi.org/10.18653/v1/D17-1263

Lambrecht, K. 1994. Information Structure and Sentence Form: Topic, Focus, and the Mental Representations of Discourse Referents. Cambridge: Cambridge University Press.

Raquel, H. 2002. Establishing topic in conversation: A contrastive study of left-dislocation in English and Spanish. Circle of Linguistics Applied to Communication, 31-50. Department of Spanish Philology, Universidad Complutense de Madrid.

Salkie, R. 2022. INTERSECT. Available at: INTERSECT Project

Westbury, J. 2016. Left dislocation: A typological overview. Stellenbosch Papers in Linguistics Plus, 50, 21-45. https://doi.org/10.5842/50-0-715

Downloads

Published

01-10-2022

How to Cite

The neural machine translation of dislocations. (2022). Linguistic Proceedings Series, 13, 137-140. https://doi.org/10.36505/ExLing-2022/13/0035/000577

Share

Similar Articles

1-10 of 399

You may also start an advanced similarity search for this article.