Research Methods in Legal Translation and Interpreting

Crossing Methodological Boundaries

Łucja Biel, Jan Engberg, Rosario Martín Ruano, Vilelmini Sosoni, Łucja Biel, Jan Engberg, Rosario Martín Ruano, Vilelmini Sosoni

• Publisher: Routledge
• Year: 2019
• ISBN:
  9781351031202
• Edition: 1
• Topic: Languages & Linguistics
• Subtopic: Translating & Interpreting
• Index: Languages & Linguistics

The field of Legal translation and interpreting has strongly expanded over recent years. As it has developed into an independent branch of Translation Studies, this book advocates for a substantiated discussion of methods and methodology, as well as knowledge about the variety of approaches actually applied in the field. It is argued that, complex and multifaceted as it is, legal translation calls for research that might cross boundaries across research approaches and disciplines in order to shed light on the many facets of this social practice. The volume addresses the challenge of methodological consolidation, triangulation and refinement. The work presents examples of the variety of theoretical approaches which have been developed in the discipline and of the methodological sophistication which is currently being called for. In this regard, by combining different perspectives, they expand our understanding of the roles played by legal translators and interpreters, who emerge as linguistic and intercultural mediators dealing with a rich variety of legal texts; as knowledge communicators and as builders of specialised knowledge; as social agents performing a socially-situated activity; as decision-makers and agents subject to and redefining power relations, and as political actors shaping legal cultures and negotiating cultural identities, as well as their own professional identity.

Chapter 2 of this book is freely available as a downloadable Open Access PDF at http://www.taylorfrancis.com under a Creative Commons Attribution-Non Commercial-No Derivatives (CC-BY-NC-ND) 4.0 license.

1 Corpus methods in legal translation studies

Gianluca Pontrandolfo

1 The perspective: key concepts

The present chapter¹ aims at studying the interaction among three conceptual areas: corpora, methods and Legal Translation Studies. More specifically, it aims at highlighting the contribution of corpus linguistics to research methods in legal translation (Biel and Engberg 2013) within the empirical and ‘technological turn’ (Cronin 2010) in Translation Studies and in line with the subfield which has been recently defined as ‘Computer-Assisted Legal Linguistics’ (CAL²) (Vogel et al. 2017).

Before delving into the analysis, some preliminary definitions of the key concepts underpinning the investigation are necessary. Firstly, within the branch of corpus linguistics, a corpus is “a collection of texts in electronic format which are processed and analysed using software specifically created for linguistic research” (Zanettin 2012, p. 7). Irrespective of its typology – monolingual or multilingual, comparable or parallel, etc. (see Zanettin 2012, pp. 10–11 for a classification applied to Translation Studies) – the common feature of a legal corpus is its “tex- tual territory” (Prieto Ramos 2014a, p. 264), i.e. the specialised nature of its texts (see section 2). Indeed, legal corpora focus on aspects related to the macro-area of law governing public or private legal relationships, applying legal instruments in specific scenarios, conveying specialised knowledge on sources of law and legal relationships (Prieto Ramos 2014a, pp. 264–265). Secondly, a method is understood here as a systematic procedure, technique, or mode of inquiry employed by or proper to a particular discipline or art. Corpus linguistics, which may be conceived as a strongly empirical methodology in itself, allows for an application of various kinds of methods that focus on specific aspects of legal language or translation (see section 3). Finally, Legal Translation Studies (LTS) is con- ceived as an interdisciplinary field that focuses on all the factors involved in the translation of legal texts, e.g. processes, products, agents, texts (Prieto Ramos 2014a, p. 261).

The objective of this chapter is to analyse the ways legal corpora have been used so far and can be used in the future to study legal translation. The angle of analysis is the research method, in line with the overall goal of this volume. The chapter will problematise and exemplify some of the most important methodological issues underlying the use of corpora in LTS. It will also review major trends of research using legal corpora. The focus will be placed specifically on studies on legal translation, rather than on legal language (e.g. Goźdź-Roszkowski 2011) or translator training (e.g. Monzó Nebot 2008), even though corpora are suitable for linguistic and training purposes as well. The pros and cons of respective methods and some examples of empirical applications will also be presented.

From a methodological point of view, it is worth underlying that the chapter is conceived as an overview of methods (various corpus approaches to legal translation) rather than an overview of the existing studies (literature review) on corpus and legal translation, which would go beyond the scope of this contribution. In fact, it would be difficult to carry out a comprehensive and up-to-date review of all corpus methods that have been and are being applied to study legal translation due to the increasing popularity they have enjoyed so far (Biel 2018a, pp. 27–28).

2 The substance: legal corpora

Attempts at classifying legal corpora may be found in the literature (see Pontrandolfo 2012; Marín Pérez and Rea Rizzo 2012; Vogel et al. 2017; Biel 2018a). The picture shows that legal corpora are increasingly growing and are used in LTS, although not all legal corpora have been explored for translation purposes.

As far as the “textual territory” is concerned, the notion of ‘legal texts’ has been widely and diversely applied in LTS. Texts pertaining to the legal sphere have been generally categorised according to different criteria, such as a nature, function or dominant text types. One of the common classifications adopted for translation purposes considers three types of texts: normative (e.g. legislative texts), interpretative (legal scholarly writings) and applied (private, administrative, judicial texts) ones (among the various classifications, see Mortara Garavelli 2001, pp. 25–34; Borja Albi 2007, p. 161; Šarčević 1997, pp. 11–12 based on Bocquet 1994, p. 2; Cao 2007, pp. 9–10; Prieto Ramos 2014a, p. 265 and Prieto Ramos, this volume).

Legal corpora may incorporate any type of these genres, but, on the practical side, LTS scholars usually face a key methodological problem, which is the question of availability and accessibility of legal texts. As pointed out by Vigier and Sánchez Ramos, despite the widespread use in other fields within Translation Studies, the development of corpora has been rather slower in the field of Legal Translation, probably due to the confidential and private nature of many legal documents (2017, p. 261). This is what Biel calls “legicentrism”, which is the tendency of existing corpora to be mainly composed of legislation (2018a, p. 29), which results in an underrepresentation of other genres.

The reality is that the majority of legal corpora are institutional ones² (see Biel 2018a, pp. 19–33): collections of texts produced, for example, at the EU, UN, or WTO,³ such as legislative or judicial texts available in many languages, often because they are supranational official versions of the same legal instrument, collected mainly with the purpose of training machine translation systems. Moreover, as pointed out by Vogel et al. (2017, p. 12), most corpora are compiled in the context of a specific research project⁴ to serve its research questions and are thus rather small and not always publicly available.

3 The avenues: corpus methods in descriptive legal translation studies

Analysing corpus methods in LTS means exploring the different paths chosen by scholars to approach translation. Corpora are obviously just one of the methodologies allowing for the study of legal translation, the others being, for example, genre analysis (Bhatia 1997), discourse analysis (Pontrandolfo 2019, forthcoming), comparative law (Engberg 2013), sociology of profession (Koskinen 2000; Lambert 2009), and functionalism (Garzone 2000) (see the Editors’ Introduction to this volume).

There are several advantages of using corpus methods in LTS. Firstly, scholars who adopt data-driven approaches have an increased methodological awareness and rigour; secondly, empirical methodologies allow for a significant shift from a prescriptive to a descriptive view on legal translation: describing “what translations actually are, rather than simply prescribing how they should be” (Pym 2014, p. 63) proves to be a key perspective in Legal Translation Studies and practice. Finally, corpus methodologies allow for methodological eclecticism (i.e. the possibility to triangulate different methods) and reduced speculation, as well as offer the potential to verify hypotheses more systematically (Biel and Engberg 2013, p. 5).

In the following sections, corpus methods in legal translation research will be analysed through the lens of typical methodological dichotomies. These methodological oppositions are a means of providing a clear picture of the trends in action in legal translation studies based on corpora, since, as it will be seen, many of them can be combined and adopted simultaneously in a single study.

3.1 Local versus global

One of the first dichotomies is the local versus global opposition. As a matter of fact, legal corpora can be used to study microstructural linguistic phenomena (local), such as terms, phrasemes, syntactic patterns, discourse markers, pragmatic features, cohesive devices, etc., or macrostructural phenomena (global), that is aspects related to, say, the genre of legal texts, such as rhetorical moves, performative macro-utterances producing legal effects, stylistic features, etc. (see Pontrandolfo 2019, forthcoming).

From a legal translation point of view, this opposition needs to be taken into consideration when choosing a method of analysis. Emphasis is often put on microstructural traits rather than on global characteristics of legal texts, which is the reason why analysing genre-related features in LTS may be problematic. The opposite situation may occur when the risk is missing the trees (single texts or local features) for the forest (the collection of texts or the global traits) (Egbert and Schnur 2018).

3.2 Quality versus quantity

One of the key debates around corpus linguistics relates to the quality versus quantity dichotomy. The quantitative dimension focuses on the importance of empirical data in confirming or rejecting hypotheses on legal translation, whereas qualitative approaches tend to privilege the centrality of discursive examples rather than (co)occurrences.

This dichotomy can also be associated with the manual versus (semi-)automatic opposition in corpus studies in that quality research seems to rely on the manual reading of legal texts whereas quantitative research tends to explore data semi-automatically with the help of a wide range of software for the analysis of linguistic data (AntConc, ParaConc, WordSmith Tools, Sketch Engine, etc.).

Corpus research is sometimes criticised due to the difficulties scholars may experience when studying more qualitative aspects of legal translation, thus limiting the results to the product rather than giving insights into the process. Indeed, quantitative results may often excessively rely on the ‘pattern’ rather than on the ‘text as a whole’ (see Egbert and Schnur 2018, pp. 159–161). This is particularly important in the legal field and in legal translation where the complexity of texts requires a macroanalysis of its context of production.

As it will be demonstrated in the next sections, the recent trend in LTS is to combine the quality and quantity dimensions in corpus methods, triangulating research methods and thus looking at the data from different perspectives.

3.3 Corpus-based versus corpus-driven

The distinction between a corpus-based and corpus-driven language study was introduced by Tognini-Bonelli (2001). Corpus-based studies typically use corpus data in order to explore a theory or hypothesis, aiming to validate, refute or refine it. Thus, they privilege an inferential approach to the analysis of data. Corpus-driven approaches are based on the idea that the corpus itself should be the sole source of our hypotheses about language (they tend to adopt an exploratory perspective on empirical data) (2001, pp. 65, 84–85). Although critical voices were raised after the distinction was established (see, for example, McEnery et al. (2006, p. 8), who consider the distinction overstated), keeping the dichotomy is useful from a methodological point of view since most of the studies inevitably combine the two approaches.

In respect of legal translation, this dichotomy proves to be an important parameter to consider when choosing the method. Corpus-based studies are particularly effective when investigating features of legal language in translation since its traits have been clearly identified in the literature (e.g. nominalisations, passive voices, lexical doublets and triplets) and can be used by researchers to test their presence in the texts with empirical data. Corpus-driven research is effective when there are no previous pre-construed ideas or expected results from the corpus analysis. Obviously, there are no pure corpus-based or -driven studies: a combination of both methods proves to be a feasible and productive perspective in LTS.

3.4 Monolingual versus multilingual

The monolingual versus multilingual dichotomy lies at the heart of all the oppositions mentioned in this section. As suggested by McEnery and Hardie, many corpora are monolingual in the sense that, while they may represent a range of varieties and genres of a particular language, they are nonetheless limited to that one language (2012, p. 18). Bilingual and multilingual corpora can use different criteria as well, such as the number of languages involved and the content or form of the corpus (2012, p. 19).

Monolingual corpora, which are frequently used in legal language studies to draw comparisons among legal genres (see, among others, Goźdź-Roszkowski 2011; Breeze 2013), are almost exclusively used in LTS to study regularities of translated language. An example would be exploring the differences between original versus translated legal texts (see section 3.6). Multilingual corpora are instead of great help when studying the translation product across languages (e.g. how a legal text is translated into a different language).

This dichotomy is closely related to the following ones: comparable versus parallel (section 3.5) and translated versus non-translated (section 3.6).

3.5 Comparable versus parallel

The comparable versus parallel dichotomy may be framed within contrastive linguistics. As Zanettin puts it, “contrastive descriptions are important to translation studies in that they provide the touchstone for assessing the extent to which typical features of translated texts and behavioural patterns of translators are determined by the source language” (2012, p. 25).

Comparable corpora can be defined as collections of original texts in more than one language or variety (the similarity may lie in the genre, proportion of text, domain, sampling period, etc.). Parallel corpora are instead collections of source texts and their translations into one or more languages.

If comparable corpora are generally the basis of translation-oriented studies aimed at searching for ‘parallel routines’ (functional equivalents) in original texts, parallel corpora are usually used to explore how a specific linguistic feature (term, verb, phraseologism, etc.) of legal languag...