ParCorFull2.0: a Parallel Corpus Annotated with Full Coreference

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 805–813 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 805 ParCorFull2.0: a Parallel Corpus Annotated with Full Coreference Ekaterina Lapshinova-Koltunski1, Pedro Ferreira2, Elina Lartaud3, Christian Hardmeier3,4 1Saarland University, 2University of Aveiro, 3Uppsala University, 4IT University of Copenhagen e.lapshinova@mx.uni-saarland.de, pedroaferreira@ua.pt, elina.aurelia@gmail.com, chrha@itu.dk Abstract In this paper, we describe ParCorFull2.0, a parallel corpus annotated with full coreference chains for multiple languages, which is an extension of the existing corpus ParCorFull (Lapshinova-Koltunski et al., 2018). Similar to the previous version, this corpus has been created to address translation of coreference across languages, a phenomenon still challenging for machine translation (MT) and other multilingual natural language processing (NLP) applications. The current version of the corpus that we present here contains not only parallel texts for the language pair English-German, but also for English-French and English-Portuguese, which are all major European languages. The new language pairs belong to the Romance languages. The addition of a new language group creates a need of extension not only in terms of texts added, but also in terms of the annotation guidelines. Both French and Portuguese contain structures not found in English and German. Moreover, Portuguese is a pro-drop language bringing even more systemic differences in the realisation of coreference into our cross-lingual resources. These differences cause problems for multilingual coreference resolution and machine translation. Our parallel corpus with full annotation of coreference will be a valuable resource with a variety of uses not only for NLP applications, but also for contrastive linguists and researchers in translation studies. Keywords: coreference, entity reference, event reference, cross-lingual coreference resolution, coreference annotation, linguistic annotation, machine translation, multilingual NLP, multilinguality, English, German, French, Portuguese 1. Introduction We present ParCorFull2.01, an extension of the existing corpus ParCorFull (Lapshinova-Koltunski et al., 2018). ParCorFull is a multilingual parallel corpus, originally containing the languages English, the original language of the included texts, and German. In this work, we describe its extension to French and Portuguese. Par- CorFull has full coreference annotation, which means that it contains not only annotation of pronouns, but also full nominal phrases, verbal phrases and clauses and includes rich set of links with both entity and event coreference. The corpus was created to study, model and evaluate the translation of coreference and coher- ence patterns in machine translation and multilingual NLP. The first version of the corpus was already used in a number of studies (Lapshinova-Koltunski et al., 2020; Lapshinova-Koltunski et al., 2019; Guillou et al., 2018). French and Portuguese, the two Romance languages added in the new version of the corpus, pose new annotation challenges as they contain structures not available in English and German. Therefore, the ex- tension of the corpus with translations into these two languages also required the revision and extension of the annotation guidelines, particularly with respect to the handling of clitics and personal pronouns in dir- ect object function, as well as relative pronouns and reflexives. Moreover, Portuguese is …

Content truncated to 500 words