EPIC UdS - Creation and Applications of a Simultaneous Interpreting Corpus

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 1193–1200 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 1193 EPIC UdS - Creation and Applications of a Simultaneous Interpreting Corpus Heike Przybyl, Ekaterina Lapshinova-Koltunski, Katrin Menzel, Stefan Fischer, Elke Teich Saarland University Campus A2.2, 66123 Saarbr¨ucken, Germany {heike.przybyl, stefan.fischer}@uni-saarland.de {e.lapshinova, k.menzel, e.teich}@mx.uni-saarland.de Abstract In this paper, we describe the creation and annotation of EPIC UdS, a multilingual corpus of simultaneous interpreting for English, German and Spanish. We give an overview of the comparable and parallel, aligned corpus variants and explore various applications of the corpus. What makes EPIC UdS relevant is that it is one of the rare interpreting corpora that includes transcripts suitable for research on more than one language pair and on interpreting with regard to German. It not only contains transcribed speeches, but also rich metadata and fine-grained linguistic annotations tailored for diverse applications across a broad range of linguistic subfields. Keywords: Simultaneous interpreting corpus, European Parliament data, multilingual language resource, corpus anno- tation 1. Introduction The European Parliament (EP) has been described as “one of the most promising sources of interpreting and intermodal corpora” for European languages due to the availability, accessibility and controlability of data and the professional interpreter status (Bernardini et al., 2018). Various interpreting corpus projects make use of this rich data source for interpreted speech. At the same time, the EP is a rather specific interpreting set- ting. EU interpreters work against the backdrop of ex- treme multilingualism in large meetings (Graves et al., 2021) and the EP context is not necessarily entirely rep- resentative for all interpreting services provided in the EU. Due to the generally time-consuming nature of spoken language corpora compilation, simultaneous interpret- ing corpora are still rare. They exist for a restricted set of language combinations, and most of them can be classified as nano-corpora (Defrancq, 2018). New interpreting corpora are currently being compiled with different goals and users in mind, and thus with vary- ing transcription levels, meta-data attributes and with different degrees of availability. These corpus projects are highly relevant as interpretese still remains a largely understudies phenomenon. We briefly discuss a non- exhaustive list of examples in this paper (Section 2). Using the standards established for existing resources, the EPIC UdS (European Parliament Interpreting Cor- pus at Saarland University) project was designed in a way that its transcriptions can be used in combination with data available within previously compiled EPIC corpora for other language combinations (cf. Section 3.2). Furthermore, we provide various corpus variants that can be used for different research foci, e.g. one corpus variant including disfluencies such as filled and silent pauses for studying cognitive aspects of interpret- ing vs. a “cleaned” corpus variant without disfluencies enabling more accurate linguistic annotations needed for research on specific lexical and syntactic features. The remainder of the paper is organized as follows: In Section 2, we provide an overview of other existing in- terpreting corpora and discuss …

Content truncated to 500 words