1 Introduction

The demand for quality has been a catalyst for both the recognition of legal translation as professional practice, particularly in multilingual contexts, and for the development of Legal Translation Studies (LTS). The debate on the translation of the Swiss Civil Code at the beginning of the 20th century, the institutionalization of translation in the emergence of the new international legal order after the Second World War, and the introduction of a system of co-drafting in Canada as a way of improving the quality of Canadian legislation in French in the 1970s provide early illustrations of that process. In these contexts, bridge-building between professional practice, research and training had its pioneers in the Geneva school of legal translation [5] and the school of jurilinguistics in the case of Canada [17].

After a few decades of expansion and academic consolidation of LTS, quality enhancement remains a major challenge in the field. As Levý [28, p. 77] pointed out when modern Translation Studies (TS) was in its infancy, “writing on the problems of translation has any sense at all only if it contributes to our knowledge of the agents which influence the translator’s work and its quality.” This assertion holds true today for research on legal translation quality (as product), and its relation to decision-making and competence-building processes and, ultimately, for the development of LTS itself [42]. Despite the proliferation of scholarly work and theoretical models, translators have been “guided more or less by their own intuition” in dealing with legal translation problems [46, p. 47], and methodological problems seem to persist even in institutional contexts where quality assurance procedures are systematically implemented (e.g. [39]).

The consequences of inadequate legal translations can be dire (see e.g. [10]) for individuals, for private or public entities, and for legal certainty in general. This explains, to a great extent, the high expectations for quality in this field, and the common perception that legal translation is a special branch of professional translation for which relevant qualifications and technical expertise are required (see overview of profiles in [6]). In this context, considering the polarization of the broader translation market with regard to quality, it is in the interest of both practitioners and scholars to prove and measure the value of translation quality from a professional perspective: what characterizes quality in legal translation, which factors have an impact on it, who can assess quality markers and how can these be enhanced.

The next sections will explore these questions by focusing on the distinctive features of legal translation, particularly its decision-making parameters and the competence required for their application, as the basis for quality evaluation. The lack of specific benchmarks in this field will first be highlighted (Sect. 2) before describing a process-based approach to quality (Sect. 3), and its implications for quality assurance, assessment and management (Sect. 4). The framework presented below is supported by observation of different professional contexts of legal translation, including institutional practices, and will be used for further empirical research on the correlation between levels of legal translation competence and levels of product quality.

Throughout this article, the term “quality assurance” will be used as defined by Saldanha and O’Brien [45, p. 95]: “systems and processes used to help create or maintain quality.” In line with this broad sense, “quality control and quality assessment are contributions to quality assurance” [33, p. 118]. In turn, defining translation quality “inevitably involves measurement and, more often than not, judgement” [29, p. 140]. While the use of this terminology may vary between authors,Footnote 1 for the purpose of this study, we will consider any actions taken to improve the quality of translations as products, regardless of the agent and the stage in the production process. These include self-revision by the translator, corrections by revisers and checks or readings by other “quality controllers” [33, p. 110], although these might not always involve entire texts or a comprehensive set of evaluative criteria. In our analysis of quality through an LTS expert lens, perceptions of lay readers and managerial considerations will be related to the ultimate goal of maximum quality from that LTS perspective, even if, as we will see in Sect. 4, from a management point of view, a fit-for-purpose translation might not necessarily be a top quality one.

2 The Need for Quality Benchmarks in Legal Translation

The first question that arises when analyzing translation quality is how to define it, i.e. what makes a quality translation, or what criteria can be used to measure quality. As noted by House [25, p. 1], “different views of translation lead to different concepts of translational quality, and hence different ways of assessing quality”. Indeed, perceptions of quality have evolved considerably since the first international conference devoted to the topic by professional translators, the 3rd Congress of the International Federation of Translators (1959). The proceedings of that conference [11] “display a puzzling array of basically vague and unverifiable statements of opinion suggesting, for instance, a connection between the quality of a translation and the personalities of the author, the translator and the audience, or assessing that a good translation is one which does not read like one” [25, p. 2]. As highlighted by Maier [29, p. 139], the proceedings also reveal a gap between literary and pragmatic translation: the “vagueness and subjectivity” of respondents to a survey on quality in the first area contrast with “a more nuanced and less subjective understanding of quality” among contributions on scientific and technical translation, “with respect to a wide variety of issues, such as translator training, improved working conditions for professional translators, terminology, and the need for increased research and collaboration among translators, particularly in the area of terminology”.

As with other issues in TS, the influence of textual, pragmatic and functional theories started to prevail in approaches to translation quality in the 1990s, moving from predominantly linguistic considerations to further emphasis on communicative, cultural and situational aspects. As expressed by Horguelin and Brunette [23, p. 13], “on est passé du savoir écrire au savoir communiquer”. Juliana House’s revision of her own landmark model between 1977 [24] and 1997 [25] is a case in point, together with more marked functionalist views by Hönig [22], Lauscher [27] or Reiss [44] at the turn of the millennium. Since then, the search for measurable evaluative criteria has attracted increasing attention in TS as evidenced by a number of journal issues and specialized collective monographs devoted to different aspects of quality, such as models of evaluation (The Translator, issue 6(2), 2000), revision (Journal of Specialised Translation, issue 8, 2007) and assessment (American Translators Association Scholarly Monograph Series, volume XIV, 2009). Hague et al. [19, p. 258] refer to a “substantial convergence” in translation quality assessment (TQA) over the past decade, and argue that this “reflects general agreement about the role of extra-textual factors such as audience and purpose, extra-textual factors which have long been basic to functionalism.”

Indeed, most academic approaches to translation quality refer to key concepts of “adequacy” (e.g. [1, 12]), “suitability” (e.g. [33]) or “appropriateness” (e.g. [3]) with regard to purpose and readership. Nonetheless, these notions are presented in very general terms and their application thus depends on the judgement of translators and revisers. For instance, in her framework for assessing translation ability, Angelelli associates top level in translation skill to “creative solutions to translation problems” [3, p. 41]. But why and when is a “creative solution” an adequate one? In fact, the problem of subjectivity persists. The “primary difficulty surrounding the issue of translation evaluation” lies in the “very fuzzy and shifting boundaries” of the notion of quality [7, p. 183]. According to Williams [51, p. xiv], this is mainly due to the fact that most quality models have been developed “with literary, advertising, and journalistic translation in mind” and, therefore, the “principles underlying them do not necessarily apply to other types of instrumental translation”. This is also pointed out by Stejskal [47, p. 291], for whom there is still “an enormous spread” between the creative features of literary translation and the more normative parameters of pragmatic translation.

As a major specialization of the latter, legal translation leaves little margin to creativity and subjectivity. It rather depends on legal conditions, comparative law and legal interpretation rules to achieve accuracy and adequacy. In a major proportion of legal translation work, this is a matter of legal consistency and legal certainty, and hence of conformity to legal sources and discourses, rather than a matter of individual preferences. In most existing approaches to quality, the role explicitly or implicitly assigned to intuition and the relativism associated with subjective judgements do not meet those priorities. As a result, defeatist positions such as the following may run contrary to the legal translator’s expert role in finding adequate solutions to legal communicative problems: “Quality is relative and absolutes of accuracy cease where the end user (i.e. client) imposes his own subjective preferences of style in TT. Standardization of quality is thus a fuzzy grey area.” [1, p. 498].Footnote 2

At microtextual level, the gap between general TQA models and legal translation methodology is further evidenced by the approach to terminological decision evaluation. As a distinctive feature of legal discourses, legal terminology is a central component of legal translation theory and practice, and hence of quality evaluation. It encapsulates the conceptual challenges of legal translation, and calls for specialized thematic and methodological competences, including comparative legal analysis. The complex layers of system-bound legal meaning and the concomitant problems of legal asymmetry explain, to a large degree, the limited suitability of most traditional lexicographical resources for legal translation, and advice against the practice of searching for off-the-shelf, all-purpose binary “equivalents” (e.g. [41]), which is common in other branches of translation. This is precisely the approach underlying TQA models that address specialized terminology components as a question of checking pre-established solutions in the relevant domain as part of the evaluation of form rather than a core component of strategic adequacy.

In spite of these shortcomings for legal translation quality evaluation, purpose-oriented approaches to translation quality have made considerable progress in providing linguistic and textual categories to be checked in revision and translation assessment in general (thematic coherence, style, cohesion, lexicon, language mechanics, etc.). Some illustrative taxonomies (including models of both TQA and revision) are summarized in Table 1. These can be useful for systematizing the evaluation of general aspects in any translation, particularly in connection with target language correctness and non-specialized content.

Table 1 Evaluation criteria in different approaches to translation quality evaluation
Full size table

Both Brunette’s [8] and Al-Qinai’s [1] proposals are the result of synthesis of previous studies, while Colina [12] deliberately reduces the number of categories in order to achieve a better balance “between theoretical sophistication and applicability” than previous models (p. 103). This is one of the main criticisms of academic approaches: they are often perceived by practitioners as too complex for applicability or even “impractical in the real world” [14, p. 190].

Meanwhile, the emergence of quality standards for the translation industry, especially for certification purposes, has also accelerated since the 1990s. The most influential quality standard issued to date for translation service providers, EN 15038:2006 [15], relies on revision to “examine the translation for its suitability for purpose”, which “shall include, as required by the project, comparison of the source and target texts for terminology consistency, register and style” (section 5.4.3). The standard specifies (only as a recommendation) that revisers “should have translating experience in the domain under consideration”. It stipulates that the translator “shall pay attention” to: terminology, grammar, lexical cohesion and phraseology, style, locale (local conventions and regional standards), formatting, and target group and purpose of the translation (section 5.4.1). It does not elaborate on parameters for measuring the “suitability” of translation decisions or the relevance of translation or revision competence, other than in vague terms. As regards terminology throughout the translation process, it refers to “compliance with specific domain and client terminology, or any other terminology provided, as well as terminology consistency throughout the whole translation” [section 5.4.1(a)]. In fact, as noted by Gouadec [18, p. 271], EN 15038:2006 is “no more than a compendium of what the prime contractor […] and the translator or translation company […] should do to contribute to quality assurance in translation, on the assumption that, if the conditions for quality assurance are met, the end-product will be of good quality.” In other words, while the standard may have an overall positive impact on the sector, it does not guarantee quality (see also [4]).

Likewise, ATM F2575-06, the American Standard for Quality Assurance in Translation issued by the American Association for Testing and Materials, is more oriented to translation service management than to product quality evaluation. Two sector-specific standards address error categories in product evaluation: J2450 Translation Quality Metric in the automotive industry, and LISA QA Model 3.1 in the localization industry. However, these are of little relevance to the development of quality models in legal translation.

As in the case of academic models, no taxonomy of evaluation criteria can be effective unless accompanied by the specialized competence to detect inaccuracies and deal with translation problems. This adds to the need for tailored quality models in legal translation that integrate relevant decision-making parameters and competence requirements to apply such parameters and to evaluate product adequacy at macrotextual and microtextual levels, including heightened attention to legal terminology as a quality marker.

3 A Holistic Approach to Quality in Legal Translation

The approach presented below aims to respond to the specificities of legal translation and overcome the limited applicability of general approaches to quality in this field. Firstly, it builds on decision-making parameters in legal translation as the basis to identify predictable criteria for quality assurance. Indeed, the components of the translation adequacy strategy are intended to provide a common framework for problem-solving, competence requirements and quality assessment. Such elements do not depend on subjective impressions but on legal and discursive conditions that should remain traceable between the translation brief and the translation product both in pre-delivery (self-)revision and in post-delivery assessment. As noted by Al-Qinai [1, p. 497], the “tendency to ignore the process of decision making lies behind the lack of objectivity in translation assessment” and, as a result, “any attempt to evaluate translations by analytic comparison of source text (ST) and target text (TT) is bound to divert away from accuracy without considering the procedures undertaken by the translator to resolve problems [21: 3].” Lauscher [27, p. 150] also notes that “[s]cholarly models of TQA could become more applicable in practice if the translation process were better integrated into the evaluation procedure and if the relative nature of the evaluation procedure itself and, hence, of any prescriptive judgement were investigated more closely”.

Secondly, by systematically linking legal, contextual, macrotextual and microtextual parameters of decision-making throughout the different stages of the translation process, and by integrating them in competence and product evaluation, our holistic approach responds to the call for an “operative model of analysis for translation evaluation that is capable of bringing together textual, contextual and functionalist criteria” [31, p. 274].

Thirdly, the core methodological framework is based on the observation of legal translation problems in a wide variety of professional settings (including institutional translation). It has been expanded since the late 1990s [36, 37, 39], and successfully tested in solving such problems and for training purposes at postgraduate level. It is therefore in line with Bush’s [9, p. 66] view that quality evaluation “should be related to a much closer analysis of the work of professional translators” of specific text types, a need recently reiterated by Drugan [14, pp. 45, 156].

3.1 Translation Process

The translation process outlined in Table 2 (adapted from [41]) marries widely-accepted functionalist principles with pragmatic considerations specific to legal translation. It is a circular process of the kind proposed by Christiane Nord [39] in her “looping model”. The search for adequacy guides the entire translation process, from the definition of the overall translation strategy to the verification of conformity of translation decisions to that strategy at the revision stage. Hence, the key challenge addressed is the interrelation of the different legal communicative variables that come into play to achieve and evaluate adequacy, and how to integrate them into a methodological framework that can be useful for the multiplicity of translation scenarios and textual genres comprised by legal translation (see overview of legal text types in [42]).

Table 2 Outline of translation process
Full size table

In the first stage, the elements of the overall translation adequacy strategy are defined through a dual process:

  • The analysis of the translation brief and the communicative situation: type and conditions of the translation (including extratextual factors, legal effects, quality requirements in whatever explicit or implicit form these might apply, e.g. guidelines within a private or public institution), and the relation between ST and TT communicative situations. The skopos (as a dynamic combination of elements shaping the TT communicative situation) is thus crucial to defining the translation strategy. However, it cannot be examined in isolation, but rather as the starting point to establish the connection between ST and TT legal functions, and the instrumental or documentary nature of the translation (see general distinction by Nord [34]).

  • The legal macro-contextualization of the translation process considering three parameters: legal systems involved (national, supranational or international legal realities), branches of law addressed, and legal text type and genre conventions to be respected. These parameters correspond, respectively, to different legal communicative conditions (or “coordinates”): linguistic and jurisdictional, thematic and normative, procedural and discursive.

In the ST analysis stage, the translator examines, among other features, the coherence, cohesion and stylistic features of the specialized text, and detects translation problems at microtextual level. It is particularly important to determine the relevance and system-specific nature of legal concepts and other discursive conventions in relation to the main legal functions of the texts, to understand legal meaning as intended by drafters, and to address any semantic problem (including ambiguity) by referring to the same legal sources and interpretation criteria as legal experts.

All these problems are then solved during reformulation. As mentioned above, the most distinctive ones are usually terminological, together with phraseological and semantic difficulties in achieving accuracy and adherence to legal discursive conventions in each context. A two-step analysis is proposed for decision-making:

  • Definition of the priorities or substrategy for microtextual adequacy (subordinated to the overall adequacy strategy): type of correspondence (more neutral, formal or functional, i.e. more or less oriented to the target or the source systems in inter-systemic translation) to be prioritized between ST and TT segments depending on: their function and relevance within the text, any legal constraints affecting such segments (e.g. regulatory framework on new legal realities), and receivers’ needs and expectations in terms of identification and comprehension of system-specific elements.

  • Analysis of acceptability (type and degree of correspondence) of the formulations that may meet microtextual priorities in the TT in order to apply the most adequate translation technique (or combination of techniques). Comparative legal and linguistic analysis can be critical for decision-making in this context, even when a particular solution is already established or expected in a particular setting, but not necessarily the most adequate one according to the relevant macrotextual and microtextual parameters.

Especially in instances of significant legal asymmetry, more than one technique might be applicable depending on situational factors, so absolute assertions about preferable techniques in legal translation are potentially flawed unless they refer to a particular problem in a specific scenario. The same applies to the use of machine translation in this field: the qualitative analysis of the different variables and layers of system-bound legal meaning constitutes a challenge for the production of acceptable drafts through machine translation (whether rule-based or statistical-based), even in more “linguistically-predictable” contexts. By the same token, the use of flexible notions of “correspondence” and “formulation” in our approach attempts to reflect the role of professional decision-making and departs from ambiguous or rigid ideas of “equivalence” and detrimental perceptions of one-to-one “equivalents” as static solutions to be systematically found in dictionaries and glossaries (which too often consist of translation proposals without sufficient contextualization or justification).

An important caveat in relation to the definition of microtextual priorities is that receivers’ needs and expectations are subordinated to the overall adequacy strategy, including legal accuracy and mandatory translation requirements with regard to specific legal purposes and legal genre conventions. For instance, legal consistency cannot be sacrificed in the name of readability whenever this might affect legal certainty in the case of a treaty or piece of legislation (i.e. legal certainty itself and legislators’ expectations prevail over stylistic considerations or lay readers’ preferences), while a sworn translation of a diploma needed for administrative validation purposes is expected to prioritize “neutral” or “conceptual” (source-oriented) formulations rather than functional (target-oriented) ones when transferring grades (i.e. with percentages and indication of pass marks rather than target-system grades), regardless of the individual client’s preferences or expectations (see e.g. [32]).

During the critical revision phase, the adequacy of the TT with regard to the relevant legal communicative situation is verified according to all the elements of the translation strategy at macrotextual and microtextual levels. The decision-making parameters considered in this stage are actually the same for self-revision by the translator and for subsequent revision or assessment by other quality controllers.

3.2 Translation Competence

As in the case of taxonomies of translation techniques, no model or grid for quality evaluation can be effective without the relevant competence to apply it to specialized text types and settings. In line with the above parameters, an integrative process-oriented approach is proposed for the development of legal translation competence [38]. It interweaves declarative and operative knowledge under five subcompetences coordinated by the key strategic or methodological one (basically, the ability to define and implement the adequacy strategy for each translation brief, and thus to carry out legal and linguistic comparative analysis, according to legal communicative variables). The other four subcompetences also integrate legal translation-specific elements: (1) communicative and textual subcompetence (including legal linguistic features and legal genre conventions in the source and target languages and systems); (2) thematic and cultural subcompetence (knowledge of the relevant legal traditions, sources and concepts); (3) instrumental subcompetence (use of specialized legal resources and related computer tools); and interpersonal and professional management subcompetence (including adherence to the relevant legal framework for translation practice and awareness of ethical principles and deontological issues in legal translation). This multicomponential model, the first one of its kind in LTS according to Piecychna [35], has been adopted as a framework for the European Commission-funded QUALETRA project,Footnote 3 which aims to produce a competence grid and training materials to support the application of quality requirements in Directive 2010/64/EU on the right to interpretation and translation in criminal proceedings.

Overall, a translator who specializes in legal translation broadly, and not just in a single context or text type, should have the ability to read with “legal expert eyes” and to achieve and evaluate adequacy adapting to the needs of specialized drafters and target readers, but also of non-specialized receivers, as appropriate, depending on the translation brief. This requires a comprehensive acquaintance with legal reasoning, rules of interpretation, legal sources and procedures, and legal discourses. However, the legal thematic component of legal translation expertise is not enough without the comparative and translation-oriented aspects put into practice through the key methodological subcompetence. Specialization in a TL legal system, while extremely relevant to legal translation, is only one element of the equation; the comparative legal and linguistic analysis necessary for decision-making may involve not only a SL legal system, but also intra-linguistic and inter-systemic comparative analysis in the case of translations for a global community of receivers with a common language. In any event, the process-oriented breakdown of competence requirements (in terms of interdisciplinary profiles and qualifications) is to be tailored to the specialization and quality assurance needs of each translation context, including institutional selection processes (see Sect. 4.2).

3.3 Translation Product

In this approach, quality evaluation of the translation product is guided by the same parameters that guide decision-making and competence requirements, as the degree of adequacy is to be measured with regard to the components of the overall translation strategy and subordinated microtextual priorities. This adequacy can be traced backwards from the textual surface of the TT or draft in a contrastive analysis with the ST and in relation to the translation brief. The verification of adequacy at macrotextual and microtextual levels includes not only specialized components (particularly legal accuracy, consistency and terminology as key quality markers), but also non-specialized content and general language elements (cohesion, register, punctuation, spelling, etc.) of any translation evaluation.

The weight of each element is determined by the adequacy strategy and the microtextual priorities themselves. For example, in the instrumental translation of a judgment (to be) issued by an international court, maximum accuracy and adherence to established terminology and formulas in the relevant legal instruments and precedents are central elements of translation strategy and thus important evaluative criteria; whereas the rigorous identification of source- and target-system judicial bodies and criminal legislation would be expected to constitute a microtextual priority in fulfilling the function of a prosecutor’s letter of request to be translated for foreign authorities. As noted in Sect. 3.1, legal systemic conditions of accuracy do not often leave much margin for stylistic enhancement from the perspective of lay readers, unless this is a priority in a text primarily addressed to them for informative purposes (e.g. a manual on the general features of a particular legal system). Overall, the (often deliberate) “obscurity” or “clarity” of the original cannot be radically or freely reversed by the translator, and the extent to which a translation can be an improvement on the original is very much conditioned by legal accuracy (i.e. the semantic implications of each potential stylistic improvement) and the different loyalties deriving from translation strategy in each scenario.

The expert evaluation of translation product quality cannot be guided by generalist impressions of fluency and readability often found in TS literature in relation to other branches of translation. For instance, according to Gouadec [18, p. 272], a translation of “fair average quality” is “correct, readable, and maybe even pleasant to read”, while a “top quality” one is “fluent, efficient, most readable, and ergonomic in that both contents and form are more than adequate on two counts”: improvement on the original and adaptation of “form and content to the particular public and destination within the particular conceptual-linguistic-cultural context of the reception and use of the translation by that public and destination.” These criteria cannot be considered of general applicability or valid guidance to professional legal translation. As examples, the instrumental translation of a legal report on the implementation of countervailing measures in the context of trade dispute settlement proceedings or the inter-systemic documentary translation of a notarial document on property transactions needed in probate proceedings have little chances of being “pleasant to read” or might not be improvements on their originals, and yet, they might be top quality translations. They should certainly not “adapt” content from the original, as might be the case with cultural constructs to be disseminated for advertising or recreation purposes.

Likewise, in a report on a debate about the quality of institutional legal translation organized by the European Commission’s Directorate-General for Translation (DGT), the same institution equates quality with ensuring “the best legal, linguistic and conceptual clarity possible” [13, p. 2]. It also highlights one of the cognitive psychology-based findings of a DGT-commissioned study on document quality control [16], according to which “‘fluency’ in a text (limited vocabulary, easy structure) determines how the content is evaluated” but “‘disfluent’ texts can be effective as they force a pause in processing information” [13, p. 2]. As we contended during the same debate, clarity and fluency have more to do with the drafting and quality of originals, which crucially condition their translation. This illustrates how LTS-informed evaluation of legal translation quality is yet to be fully acknowledged by those who promote quality on the basis of (partial) approaches from adjacent fields or generalist views prevalent in the “translation industry”, rather than on the basis of legal translation methodology and adequacy parameters.

4 Implications for Legal Translation Quality Assurance, Assessment and Management

Building on a process-based approach to quality, the establishment of quality assurance mechanisms and implementation monitoring by translation service providers are a matter of fine-tuning quality enhancement measures to specific needs in two directions: firstly, ensuring that the translation job is assigned to a translator with the suitable competence to provide a product of the highest possible quality, and secondly, deciding on pre-delivery revision or post-delivery assessment tasks by qualified evaluators in order to measure and, where appropriate, improve the adequacy of translation decisions. As noted above, the predictability of decision-making parameters enables better tracking of the application of translation methods and of their suitability to the relevant legal communication situation from self-revision (a form of evaluation in itself) to subsequent quality control. Different purposes and organizational structures will support different procedures and guidelines for quality enhancement considering extratextual constraints (e.g. time, resources, translators’ availability, etc.). We will briefly focus on the implications that come into play in applying such procedures in legal translation: quality metrics, evaluators’ competence, and management challenges.

4.1 Quality Levels

A variety of quality metrics systems would be compatible with the holistic approach presented above as long as they respect the variables of legal translation methodology, i.e. if they articulate predictable evaluative criteria based on decision-making parameters, and provided that quality evaluators have the relevant expertise to assess such criteria (see Sect. 4.2). The exact definition and weighting of those criteria and the categorization of different levels of performance would depend on the particularities of each context, notably the scope and purpose of the evaluation. For example, the assessment priorities would vary between: (a) the screening of freelance sworn translators for a transnational law firm with a diversity of legal translation needs, and (b) the recruitment or appraisal of translators within a hierarchical institutional structure dealing with a specific set of legal texts on a particular branch of international law. Furthermore, as proven in practice, the same parameters can be extrapolated to assessment in translator training for such professional contexts.

A possible template of core evaluative variables in legal translation would include three broad categories for quality measurement:

  • Legal semantic accuracy and legal consistency.

  • Adequacy of translation decisions on legal discursive features (including legal terminology, legal phraseology and legal genre conventions) considering the overall translation strategy and the microtextual priorities (see Sect. 3.1).

  • General linguistic correctness (including cohesion, syntax, punctuation, etc.).

The resulting levels could be expressed with a variety of taxonomies, including as many categories as deemed appropriate. Five grades are broadly described in Table 3, but these could be reduced or expanded, and the different labels could be adapted to different assessment needs. For example, an “unacceptable” translation could correspond to level E or level 1 in an A-to-E or a 5-to-1 descending scale, respectively.

Table 3 A general description of quality levels in legal translation
Full size table

Each general category would integrate scores on different components of evaluation to be defined and coherently weighted in each assessment exercise, i.e. the numerical values and thresholds for each component, whether marked with an error-deduction approach, a non-punitive approach or a combination of both (see e.g. [48] and [50]). In our view, as captured in Table 3, it is mandatory to carefully analyze the type and severity of content inaccuracies, linguistic errors and inadequate methodological decisions on legal terminology and discourse, even if these are contextualized within holistic descriptors in which adequate solutions are also valued. As anticipated in Sect. 2, general models of translation quality evaluation can be integrated or made compatible with this kind of overarching benchmark (see Table 1 and the American Translators Association’s rubric for grading, [2]).

The flexibility to choose and adapt marking systems does not entail a distortion of the real degrees of adequacy from an LTS perspective. It rather means that these levels are managed in relation to different expectations and uses in assessment. For example, level 1 would be a failed attempt at communication and an unacceptable translation by all means; a poor translation of level 2 might be enough to get the gist of a document for basic informal needs, but clearly inadequate for legal purposes; while the grade assigned to level 3 in a training context may correspond to a pass or fail value depending on the difficulty and expectations of each course and test, although the same level 3 would be insufficient for professional accreditation (and trainees should be aware of it). As argued by Lauscher [27, p. 164], TQA is ultimately “a matter of communication, co-operation and consent”, and this applies to an approach in which, regardless of the marking system agreed by evaluators, the core evaluative criteria remain predictable and objective on the basis of shared decision-making parameters in legal translation.

The intertwining between process, competence and product also makes it possible to associate different levels of legal translation (sub)competence to different degrees and components of adequacy in evaluation.

4.2 Quality Evaluators

The higher the quality expected in legal translation, the higher the expertise required not only to achieve adequacy but also to evaluate it. The components of legal translation competence outlined in Sect. 3.2 obviously apply to revisers and other quality controllers, whose interdisciplinary expertise is normally expected to be higher than (or at least as high as) that of the translator in order to detect translation problems and improve and/or assess their solutions. This, however, is not the norm in practice. In the case of pre-delivery quality control, input by a reviser is a general precautionary measure, but not a guarantee of maximum quality, as illustrated e.g. by Künzli [26] and Martin [30]. The reviser’s specialized competence and extratextual constraints (see Sect. 4.3) are key factors in the equation. Four eyes are normally more effective than two if they are all qualified professionals with the right expertise and working in optimal conditions. When this is not the case, a single qualified legal translator might produce a better translation than a combination of translator and reviser.

The same logic applies to subject matter specialists with no expertise in legal translation. Their thematic competence can be of high value to the translator, particularly any advice on legal concepts and discourses, but they usually lack the contrastive linguistic and methodological tools to replace a translation expert in the evaluation of translation decision-making and product. Nonetheless, when thematic specialists are priority target readers and/or clients (e.g. delegates or legislators at international bodies) and discrepancies arise between these and translators, the latter must try to strike the right balances between different kinds of input and “loyalties” for the sake of quality. This might involve some pedagogical effort by the translator in negotiating instructions or preferences received from clients. As Mossop ([33], p. 23) reminds us, “most clients know next to nothing about what translation involves”, which supports an obvious maxim often ignored by outsiders to professional translation (and by many translators themselves): “In no profession can one always bow to the client’s wishes”. This would equate, for instance, to a patient telling a doctor how to deal with a medical problem rather than just explaining the symptoms. Why then resort to professional help instead of homemade remedies?

Needless to say, the relevance of evaluation by clients or readers who are not technical experts is even more limited, as is their competence to judge legal translations. Lay readers’ judgements may focus on TT readability or suitability for general public understanding, but, as noted in Sect. 2, this is secondary to other elements of the adequacy strategy in most legal translation scenarios. Otherwise, a non-expert TT evaluation is more often than not a subjective statement on the complexity of law or legal discourses, rather than on the quality of legal translation. In sum, translators cannot overestimate “client satisfaction” in instances where the client has insufficient knowledge of the original language, the subject matter and translation methodology so as to be able to fully appreciate the translation product.

Last but not least, trainers hold the responsibility of keeping their specialized competence up to the level required for the evaluation of trainees in line with course and professional expectations. This integral approach presupposes a continuum between formal training and professional practice (see [38]) in which trainers can only be effective evaluators (at least, in training for “real-world” situations) if they are up-to-date with quality standards in the settings for which they instruct students. As in any other field, a trainer who is unable to provide a top quality product would hardly be in a position to dissect all the nuances of problem-solving necessary to reach that level of quality, even if s/he might be able to facilitate and assess progress in that direction.

The question of who can evaluate each aspect of legal translation quality in each scenario is related to the issue of who certifies the certifiers in a broad sense. This issue is common to other areas, but particularly sensitive in young (and booming) fields like LTS (and TS in general). The limited number of qualified experts in evaluation panels can be a matter of concern in some constituencies. Echoing this concern, a recent report [20] prepared for the Australian National Accreditation Authority for Translators and Interpreters (NAATI) has recommended that the composition of its examiners’ panels be reviewed “to include more graduates of approved courses and fewer practitioners who hold no formal qualifications in Interpreting and Translation” [20, p. 89]. The same question could be posed in connection with the competence requirements (qualifications and proven translation skills) for the screening of revisers, examiners, trainers or evaluators in other contexts.

4.3 Quality Managers

As put by Vlachopoulos [49, p. 17] in relation to the DGT, “the improvement of translation quality is as much a managerial challenge as it is a linguistic and technical one.” Indeed, in any public or private organization with a structured translation service, managers play the key role of promoting quality across the board (including in translation professionals’ recruitment, task assignment, monitoring and appraisal exercises, training policies and interaction with clients or users) and addressing all extratextual factors and practical constraints (such as deadlines, tools used and remuneration) with a potential impact on the workflow and product quality in each particular translation job. Overall, the implementation of quality assurance policies entails matching offer and demand in a cost-effective way, that is, in order to achieve the highest possible quality with the available material and human resources, and within the relevant timeframe.

This implies that, considering budgetary and time limitations, top quality might not be affordable or possible to guarantee at any given time. Working on this presumption, large translation services at international organizations (see e.g. the case of the OECD in [43]) tend to apply quality assurance procedures in which texts are assigned to different translators, and subject to varying levels of revision, depending on the importance of the document. In this context, legal instruments and binding decisions (mostly legislative and judicial genres) are found at the top of the institutional text hierarchy, while documents on legal implementation or monitoring procedures usually rank lower. In a multilingual legal setting, assuring the highest quality of instrumental translation involves ensuring not only consistency and accuracy per language pair, but also maximum interlinguistic concordance between all the authentic languages (see [40]). In the case of private organizations, the systematization of quality control practices tend to be more erratic, as economic pressures may override quality considerations for the sake of profit-making or budget constraints.

In any event, a distinction must be made between fit-for-purpose managerial decisions (reasonable solutions depending on quality needs, time and resources) and adequate translation decisions required for a top quality product (see Sect. 4.1). It also becomes apparent that long-term quality enhancement policies require the evaluation of quality assessment and management practices themselves. Given the influential role of translation quality managers, high translation competence would be expected of them, together with sound managerial skills, especially in small teams where the manager becomes the only in-house quality filter. While this applies to all translation branches and will not be further examined here, it must be reminded because quality management can have a strong bearing on the implementation of holistic approaches to quality.

5 Conclusions

The quest for quality in legal translation has always been a central remit for LTS. The interest to prove and safeguard the value of quality is gaining momentum in an increasingly fragmented translation market. General approaches to translation quality fall short of the need for benchmarks aligned to the communicative conditions of legal translation. Against this background, the holistic approach outlined above integrates contextual, macrotextual and microtextual variables of decision-making, as well as concomitant competence requirements, for achieving and evaluating the adequacy of legal translation product.

The use of expert decision-making parameters as a yardstick for legal translation quality enhancement can only reinforce the predictability and objectivity of evaluative criteria (including semantic accuracy, legal consistency and adequacy of legal discursive features) and the traceability of translation problem-solving, from the revision of the translation product (or draft) backwards to the stages of translation strategy definition, text analysis and reformulation. Legal translation quality assurance is accordingly understood as an activity carried out by experts capable of measuring the adequacy of translation decisions in relation to a given translation brief and communicative situation, relying on commonly-agreed decision-making parameters and evaluation criteria. Quality levels are presented as degrees of adequacy, avoiding static perceptions of translation that run contrary to tailored methodologies for legal translation practice. In our overarching model, the metrics of the core quality components can be adapted to specific scenarios and evaluation purposes, in a continuum between training and professional practice. Agreement on this kind of framework can thus help to overcome traditional risks of subjectivity in TQA.

This type of holistic “surgical approach” fits into an empowering paradigm of legal translation as an expert activity, and serves to debunk simplistic perceptions of TT quality, particularly among non-experts in translation. It entails an important effort of systematization of legal translation variables, as well as questioning partial quality assurance practices. In this respect, it would be useful to agree on a comprehensive quality benchmark linking product, process and competence, along the lines of the Common European Framework of Reference for Languages (CEFR). However, this is still far from sight considering the issues at stake, especially if a metric standard was to be consistently applied for professional accreditation and training programme evaluation on the basis of real evidence of translation quality (rather than the wishful thinking found in the EN 15038:2006 standard and parallel training initiatives).

In spite of increasing convergence in the field, the diversity of expectations and approaches and the lack of quality control in a large proportion of translation services make the issue of quality grades and certification a sensitive one, and less straightforward than, for instance, assigning energy efficiency labels. Nonetheless, LTS can contribute to raising quality standards by proposing sound methodological models and measuring their impact on the translation product, especially in contexts where quality is genuinely valued. The relevance of LTS to professional practice would be strengthened as a result.