Corpus-Based Interpreting Studies

Following a number of studies on translator style, this paper attempts to apply corpus linguistics methods used in translation studies and inspired by stylometry to verify whether interpreters also demonstrate features of individual style.
Interpreting style is viewed here as a way of interpreting which is characterized by a consistent prominence of one or many linguistic features in interpretations from different source languages that make one’s interpretations distinguishable from those of others, as well as from one’s own speaking style. Style is approached from the point of view of authorship attribution and stylometry, i.e. described by quantitative linguistic discriminators, which in the present study are the parameters indicating the tendency to repetitiveness, informativeness and lexical
sophistication.

The paper introduces a corpus of simultaneous interpretation, SIREN. SIREN is a parallel aligned bidirectional corpus of original and simultaneously interpreted speech in Russian and English. At the moment the corpus contains 235,040 words and is enriched with POS and shallow syntactic annotation. After outlining the corpus design, I used scores for lexical variety, density and POS proportionalities to make tentative claims about the linguistic variation between originals and interpretations. Low lexical variety and density are taken as indicators of simplification, while a higher ratio of nominal to pronominal reference is seen as an indicator of explicitation. Atypical word-class distribution indicates the source language shining through. Somewhat contradictory results, with the Russian subcorpus conforming to the predictions of translation theory and the English subcorpus exhibiting the opposite trend in all universals but still shining through, invites further investigation of the data and once again puts into question unequivocal claims about T-universals.

The chapter outlines a corpus-based analysis of topical coherence in interpreted American presidential debates broadcast on Italian television. It aims at investigating the ways in which dialogue format and question/answer structure are managed. The first part identifies the types of question and answer in the SL, their Italian equivalent, and the incidence of each type of question. The question/answer classification takes into account syntactical, illocutionary and perlocutionary aspects of the discourse. The second part focuses on question/answer topical coherence in the interpreted versions. It examines whether topical coherence is achieved, and in which ways its achievement is influenced by the type of question and the changes that occurred during the interpretation process, observed through a contrastive analysis of the original and the interpreted version.

Following a number of studies on translator style, this paper attempts to apply corpus linguistics methods used in translation studies and inspired by stylometry to verify whether interpreters also demonstrate features of individual style. Interpreting style is viewed here as a way of interpreting which is characterized by a consistent prominence of one or many linguistic features in interpretations from different source languages that make one’s interpretations distinguishable from those of others, as well as from one’s own speaking style. Style is approached from the point of view of authorship attribution and stylometry, i.e. described by quantitative linguistic discriminators, which in the present study are the parameters indicating the tendency to repetitiveness, informativeness and lexical sophistication.

For the past couple of decades, researchers of Corpus-based Translation Studies (CTS) and Corpus-based Interpreting Studies (CIS) have intensified their efforts to compile authentic data needed for empirical research (Fantinuoli & Zanettin, 2015; Hu, 2016; Russo, Bendazzoli, & Defrancq, 2018). However, most of these corpora use English either as the source/target language or as the pivot language, with the Oslo Multilingual Corpus for CTS and the European Parliament Interpreting Corpus (EPIC) for CIS being notable exceptions (ibid.). Even if dependence on English is justifiable from a practical and economical point of view, it may consequently inhibit discovery of unique phenomena among less-studied languages and language pairs, especially when non-European languages are involved.
In order to examine such a possibility, this study analyzes the outcomes from an ongoing, multi-institutional corpus project which aims to expand an existing parallel corpus between Japanese and English  to include the two most spoken languages other than English—Chinese and Spanish—using Japanese as the source language. As a pilot study, both professional interpreters and student interpreters were asked to perform simultaneous interpretation from Japanese to Chinese and Spanish that was then compared with simultaneous interpreting from Japanese to English using ELAN, an annotation software. In the analysis, linguistic interference in simultaneous interpreting as defined by Lamberger-Felber & Schneider (2008) was frequently observed, but the types were different from those typically seen in the Japanese-English language pair (homographs in the case of the Japanese-Chinese language pair and word order in the case of the Japanese-Spanish language pair). By presenting the initial results of the pilot study, this paper aims to provide comparative data to test some of the key findings of prior research regarding linguistic interference.

Key words: parallel corpus, interference, Japanese, Chinese, Spanish

References
Fantinuoli, C., & Zanettin, F. (Eds.). (2015). New directions in corpus-based translation studies. Berlin: Language Science Press.
Hu, K. (2016). Introducing corpus-based translation studies. Shanghai, China:  Shanghai Jiao Tong University Press and Berlin, Heidelberg, Germany: Springer-Verlag.
Lamberger-Felber, H., & Schneider, J. (2008). Linguistic interference in simultaneous interpreting with text: A case study. In G. Hansen, A. Chesterman, & H. Gerzymisch-Arbogast (Eds.), Efforts and models in interpreting and translation research: A tribute to Daniel Gile (pp. 215–236). Amsterdam: John Benjamins.
Russo, M., Bendazzoli, C., & Defrancq, B. (Eds.). (2018). Making way in corpus-based interpreting studies. Singapore: Springer.
Shlesinger, M. (1998). Corpus-based interpreting studies as an offshoot of corpus-based translation studies. Meta, 43(4), 486–493.

Resumen La investigación en el campo de la evaluación de la calidad en interpretación simultánea puede enriquecerse significativamente con las ventajas que aporta el uso de un corpus estructurado. En esta contribución presentamos las razones que subyacen a la creación del corpus oral ECIS, sus características y técnicas de aplicación que bien pudieran resultar útiles para otros campos. Palabras clave: corpus oral, interpretación simultánea, evaluación de la calidad Abstract Research in simultaneous interpreting quality assessment could draw significant benefits from a structured corpus-assisted approach. In this paper we discuss the reasons behind the creation of ECIS-an oral corpus-, its characteristics and some implementation techniques that might well prove useful in other fields.

his paper describes the CorIT (Corpus of Television Interpreting) project by Straniero Sergio & Falbo (Straniero Sergio, 2007; Falbo, 2009, 2012; Straniero Sergio & Falbo, 2012) and gives an account of a series of studies on television interpreting conducted on CorIT data, both by analysts and MA students of the SSLMIT of the University of Trieste. After a brief introduction (cf. 1), an illustration of Corpus-based interpreting studies (CIS) and the relevant methodological framework is provided (cf. 2), with special attention devoted to the latest developments present in the literature. In the following section (cf. 3) the focus shifts to CorIT, its main features and classification criteria. The final section (cf. 4) illustrates a series of investigations conducted on CorIT, from the point of view of the main research perspectives and research methods applied to CorIT data. By firstly illustrating the data collected in CorIT and the applied classification criteria, and subsequently providing an account of the broad range of issues investigated within the project, the paper aims at highlighting the great potential of research conducted on television interpreting performances with the corpus-based approach, especially from the point of view of data comparability and ecological validity.
OpenstarTs: http://hdl.handle.net/10077/10607

Abstract for ATISA IX: Contexts of Translation and Interpreting

Title: Exploring Corpus-based Interpreting Studies in a Non-European Context: A Japanese Case Study

In the past decade, Corpus-based Interpreting Studies (Shlesinger, 1998) has attracted increased attention in Europe, mainly thanks to the compilation of the European Parliament Interpreting Corpus (EPIC) in 2005. EPIC gave rise to a number of other projects in Europe (Straniero Sergio & Falbo, 2012), but those dealing with non-European languages are still relatively limited, with the Chinese-English Conference Interpreting Corpus (CECIC) being a notable exception (Hu & Tao, 2013). The difficulty of collecting authentic interpreting data as well as the time, money and effort required to carry out transcription, word/sentence alignment and annotation have posed significant challenges to researchers in Interpreting Studies who have become increasingly aware of the importance of being able to examine their findings quantitatively. As a consequence, large-scale interpreting corpora remain underdeveloped in many parts of the world even compared to corpora in Translation Studies, not to mention Linguistics.
This paper introduces a four-year, government-funded project to compile a large-scale parallel corpus between Japanese and English. The corpus utilizes over 300 hours of speech data obtained from interpreter-mediated press conferences held at the Japan National Press Club (JNPC) since October 2009. Being the only national press club in Japan, JNPC attracts high-level speakers including heads of states and governments, whose speeches as well as answers to questions from the journalists are videotaped along with the interpreters’ renderings. The dataset, officially made available to the project team under a Memorandum of Understanding (MOU) signed between the team and JNPC, includes both simultaneous and consecutive interpreting by a number of highly trained professionals who are identified by name and affiliation. Both the quantity and the quality of the original speech content and the interpreters’ performance, as well as the authenticity and the transparency of the dataset make the corpora one-of-its-kind between Japanese and English.
Collaborating with researchers in Europe, a seven-member project team has experimented with ways to semi-automatically transcribe these video and audio materials into texts using IBM Watson Automatic Speech Recognition (ASR).  ASR errors are manually corrected with ELAN.  ELAN is a free software which enables video and audio resources to be combined with textual data on a single screen. The translations are further aligned on sentence/word-level using a word alignment tool named YAWAT  and integrated into the CRITT TPR-DB.  The TPR-DB toolkit allows for a number of statistical analyses of the data, including measuring and examining the Ear-Voice-Span (EVS) and other dynamic factors of importance to interpreting. Throughout this pilot phase, the team has encountered both technical and methodological challenges, some of which derive from lexical differences between Japanese and English (e.g. subjects are often omitted in Japanese sentences and the verb comes last which makes the alignment process more complicated than those between European languages). As the project reaches its halfway point, it is hoped that this paper will provide new topics for discussion on the possibilities and limitations of Corpus-based Interpreting Studies as well as implications for future research, especially among scholars working with non-European languages.

Key words: interpreting, parallel corpus, press conference, Japanese, English

References
Hu, K., & Tao, Q. (2013). The Chinese-English Conference Interpreting Corpus: Uses and limitations. Meta, 58, 626–642.
Shlesinger, M. (1998). Corpus-based Interpreting Studies as an offshoot of Corpus-based Translation Studies. Meta, 434, 486–493.
Straniero Sergio, F., & Falbo, C. (Eds.). (2012). Breaking ground in Corpus-based Interpreting Studies. Bern/New York: Peter Lang.

This paper examines the handling of complex noun phrases in simultaneous interpretation into Italian of English speeches in the electronic corpus EPIC (European Parliament Interpreting Corpus). The complex noun phrases analysed in this study are noun phrases with two or more premodifying items included in the following categories: nouns, adjectives, adverbs, cardinal numbers and genitives. The aim is to extract complex noun phrases from a large sample of authentic English speeches and compare them with their corresponding translation into Italian in order to study the strategies used by interpreters. The initial hypothesis was that complex noun phrases pose a translation challenge in simultaneous interpreting from English into Italian because of structural and lexical diversities and memory overload. This hypothesis was partially confirmed in that strings where information was changed or deleted represent 45 per cent of the cases. In most cases, however, interpreters were able to adopt effective translation strategies.

Drawing on Miriam Shlesinger’s seminal paper on the creation and use of corpora in Interpreting Studies research, which she referred to as an offshoot of corpus-based translation studies (Shlesinger 1998), and on Setton’s chapter presenting an overview and prospects of CIS (Setton 2011), this chapter takes stock of nearly two decades of CIS and looks at the extent to which this offshoot has flourished in three areas of interpreting, namely research, education, and professional practice. Although considerable progress has been made in dealing with many of the challenges involved in corpus development, each area has reached a different degree of maturity with respect to CIS. In fact, corpora are increasingly used in research, not only to look at simultaneous conference interpreting, but also to investigate other modes and settings, including consecutive interpreting and dialogue interpreting. This applies to both electronic, machine-readable corpora and more traditional datasets that are analyzed manually. On the other hand, their exploitation in interpreter education is still limited, while professional practice has yet to discover the potential of corpora. A better understanding of the reasons behind these discrepancies may be helpful to inform future directions of CIS and narrow the gap between research and (educational and professional) practice. Finally, the features of Web 2.0 are considered to discuss possible solutions to some of the methodological obstacles in the creation and use of interpreting corpora.