Quality Assurance in Community Translation

Mustapha Taibi

1. Introduction

Translation and interpreting literature abounds with references to quality and effectiveness (e.g. Drugan, 2013; Hale et al., 2009; House, 1977; Moser-Mercer, 1996; Williams, 2004). However, the notion of quality in these language services is far from consensual. Understandably, each translation theory or approach defines quality differently and places more emphasis on some criteria than others. In this regard, two broad areas of literature can be identified: one academic, the other professionally oriented (Taibi & Ozolins, 2016: 108–110). The first aims to establish criteria and standards based on an academic or theoretical understanding of translation and a linguistic and textual comparison of texts. Through these criteria, target texts can be compared to source texts and a quality rating can be determined. Examples of authors representing this area of literature are House (1977, 1997, 2001, 2013 and 2015) Brunette (2000) and Depraetere (2011), whose main interest is theoretical and related to translation critique, and Petersen (1996) and Hague et al. (2011), who are more concerned with pedagogic applications of quality, i.e. translation learning, assessment and feedback to assist and facilitate development of translation skills. The other area of translation quality literature focuses more on professional production processes, i.e. translation as a process of service provision and project management (Drugan, 2013; Dunne, 2011; Orsted, 2001; Samuelsson-Brown, 2006). As some authors have indicated (e.g. Drugan, 2013; Lauscher, 2000; see also Townsley in this volume), there is a perceived gap between these two strands of translation quality. However, they – in principle – should inform and complement each other: translation quality cannot be understood in depth without a sound theoretical understanding of translation and the features of effective translation, on the one hand, and adequate knowledge of professional processes and industrial relations that operate on the ground, on the other.

Starting from this premise of interdependence and complementariness, in this chapter translation quality assurance is addressed in a specific and special subfield of translation, community translation, also known as public service translation. The first question that the title of the chapter (quality assurance in community translation) is likely to trigger is whether translation standards and quality assurance processes vary from one field of translation to another, i.e. whether community translation needs to have quality criteria and processes that are different from those applied in other fields of translation. The answer to this question is that, although the core of translation theories, assessment criteria, professional standards and quality assurance processes may apply to different types and settings of translation, each type and setting might require different or more specific considerations and applications, or might need quality processes and evaluations to place more emphasis on some aspects than others.

What distinguishes community translation from other types and domains of translation is that its main mission is to empower local communities and give their members voice and access to information, services and participation (Lesch, 1999 and this volume; Taibi, 2011; Taibi & Ozolins, 2016). Because community translation is intended to empower disempowered social groups by enabling them to have equitable access to public service information and to participation in their society, this overarching mission needs to be an essential consideration in understanding and applying quality standards in this subfield of translation and social services. Although many of the established criteria (e.g. accuracy, appropriateness, readability) and processes (e.g. selection of personnel, revision and editing) of quality assurance in translation are relevant and applicable to community translation, the nature of the latter and the specificity of the public it serves make it necessary to highlight a number of specific considerations, which go from translator recruitment to processing of text contents and treatment of the translation process itself.

Quality assurance in community translation is multi-faceted and, as in other fields, involves a number of stages, actors and actions (adequate training, appropriate recruitment processes, assessment and processing of source texts, production processes, consultation with target communities, etc.). However, as stated above, the nature of this language service gives a particular nuance to all these aspects. In the following sections I start with a brief discussion of some of the main quality issues in translation in general and, subsequently, propose a comprehensive framework for quality assurance in community translation, which encompasses not only the translation phase but also the phases preceding and following it, and not only the work of translators but also the role of other stakeholders.

2. Translation Quality

In an attempt to offer an all-encompassing definition for translation quality, Koby et al. (2014) provide both a broader definition and a narrower one.

• A quality translation demonstrates accuracy and fluency required for the audience and purpose and complies with all other specifications negotiated between the requester and provider, taking into account end user needs. (2014: 416)

• A high-quality translation is one in which the message embodied in the source text is transferred completely into the target text, including denotation, connotation, nuance and style, and the target text is written in the target language using correct grammar and word order, to produce a culturally appropriate text that, in most cases, reads as if originally written by a native speaker of the target language for readers in the target culture. (2014: 416–417)

Although quite comprehensive, well thought out and seemingly encapsulating decades-long knowledge advances in translation studies, these definitions illustrate how any aspect of translation quality may be controversial. Questions that may arise include: (1) Why are only accuracy and fluency included in the broad definition?; (2) How can the needs of end users be determined and accommodated?; (3) What happens if the requester's specifications clash with the end user's needs, as understood by the translator or another stakeholder? Is the style or register of the original an aspect that must always be mirrored, regardless what the communicative situation is?

Concern for quality in translation dates back centuries (e.g. in the case of religious and literary texts) or, at least, decades (e.g. for pragmatic or instrumental texts), as Williams (2004: xiii) observes. However, as the same author notes, although there is general agreement that translations must meet quality standards, there are disparities of views on the notions of quality, acceptability and the criteria to determine them (Williams, 2004: xiv). A number of approaches, models and quality standards have been put forward, both in translation studies (academia) and the translation industry (professional practice). Yet, none of them is able to claim applicability across text types and genres or a level of clarity and detail as to rule out assessor subjectivity.

A prominent example of the approaches in this area is the work of House (1977, 1997, 2001, 2013 and 2015), who over decades has developed a well-argued functional basis for translation quality assessment. House argues that, for the quality of a translation to be assessed appropriately, some key parameters need to be scrutinised to construct a profile of the source text and then compare it to the target text. These parameters are basically Halliday's ideational, interpersonal and textual metafunctions and their corresponding discourse descriptors: field, tenor and mode. Field `refers to the nature of the social action that is taking place' (House, 1997: 108), that is the subject area or content the text covers. Tenor `refers to who is taking part, to the nature of the participants, the addresser and the addressees, and the relationship between them in terms of social power and social distance .' (House, 1997: 108–109). Mode denotes the channel of communication (e.g. spoken vs written, as well as combinations such as written to be spoken, etc.) and the extent to which interlocutors are allowed to participate. For House, a quality translation is one that is recognisable as equivalent in terms of ideational and interpersonal meanings (field and tenor) and takes into consideration the means of communication and the level and type of participation (mode). House's parameters also include genre, a key aspect that assists in constructing text profiles and distinguishing one text from another. Although the register categories (field, tenor and mode) are useful in describing the relationship between text and context, the notion of genre serves as a means to situating texts in their deeper and broader intertextual context, i.e. in the class of text with which they have something in common, and the `"macro-context" of the linguistic and cultural community in which the text is embedded' (House, 2015: 64). Once a source text profile has been established using these textual and intertextual criteria, the quality of the target text can be assessed in light of and by comparison with this profile.

Another influential contribution to understanding and assessing translation quality is Skopos Theory (Reiss & Vermeer, 1984; Vermeer, 1989), which no longer seeks equivalence as much as it does adequacy or appropriateness. The theory presents translation as a purposeful action whose ultimate goal is adequacy between a translation and its Skopos, i.e. the intended function of the text in its new context of use. While linguistic and pragmatic models of translation have focused on equivalence for decades, Skopos Theory `dethrones' the source text and considers it as a mere `offer of information'. As such, what counts most is not fidelity to the original text, but the internal coherence of the translatum (target text). The `coherence rule' in this theory establishes that the target text must be coherent with the situation of its audience (Reiss & Vermeer, 1984: 113). In other words, it needs to make sense and be sufficiently coherent in its new context of use, and be in line with the expectations and background knowledge of the intended audience. As Nord (1997: 29) notes, the Skopos rule allows translators to determine what is best for a given text and communicative situation: whether a formally faithful translation or a free version of the original text, or any position between the two is deemed situationally appropriate. Indeed, Reiss and Vermeer (1984) do not consider that a `faithful and complete' reflection of the source text is the only valid translational option; rather, any other rendering, including summary translation, free translation or adaptation could be acceptable if appropriate and effective in a given context. In terms of quality assessment, Skopos Theory broadens the scope and understanding of translation and offers key criteria that may assist in determining quality, including translation brief, text function, adaptation to target audience and internal coherence. However, as House (2015: 11) notes, the theory falls short in terms of explicitness and operationalisation, which makes it of little use for the practice of translation quality assessment.

On the other hand, quality standards and assessment practices which have gone a long way towards explicitness, detail and operationalisation have been criticised for being too microtextual and focusing mainly on error analysis at sentence or sub-sentence level (Williams, 2004: 3–9). In Australia, for instance, the translation assessment standards of the National Accreditation Authority for Translators and Interpreters (NAATI), applicable in both its own certification testing and NAATI-approved training courses, are based on point deduction for mostly accuracy and language errors. Candidates sitting for the Professional Translator level are expected to translate two 250-word texts and obtain a minimum mark of 70/100. Examiners are provided with guidelines and examples of errors and the number of points to be deducted for each error depending on its weight (Turner et al., 2010: 14). NAATI also provides a list of recommended symbols to be used when marking translations. This is intended to ensure consistency among assessors and feedback clarity for candidates. Errors are classified `in terms of accuracy, quality of language and technique' (Turner et al., 2010: 14). Accuracy and language appropriateness are dominant, though, as errors are further categorised as `general errors' (serious mistranslations or major omissions), `isolated errors affecting accuracy' (e.g. mistranslation of a given word, unjustified addition, wrong word order, grammatical error, etc.) or `isolated errors not affecting accuracy' (e.g. unidiomatic word choice, superfluous lexical items, spelling, etc.).

Although the NAATI model dates back earlier, the adoption of quantifiable error types characterised a shift towards more `objective' translation quality assessment in different parts of the world in the last half of the 1990s and the beginning of the 21st century (Lommel et al., 2014: 457). This came as globalisation was becoming a buzzword, compliance with ISO quality standards was gaining ground as a performance indicator in industries, and translation was coming of age as a language industry (Williams, 2004: xiv). Probably the most recent development in this shift is the Multidimensional Quality Metrics (MQM) system (Lommel et al., 2014, 2015), which was developed as part of a European Union-funded project and aims to offer an industry-wide acceptable framework within which all types of translation can be evaluated as `objectively' as feasible. The framework does not put forward a one-solution-fits-all approach, but offers a comprehensive list of over a hundred translation and production issue types, from which stakeholders can choose the most relevant to the assessment exercise at hand (Lommel et al., 2014, 2015). The issue types included in MQM fall under well-known translation dimensions such as accuracy, fluency, style and terminology, but also cover aspects that are relatively new, at least in terms of labelling, such as design (i.e. format or presentation of a text), internationalisation (i.e. work on the source content to make it ready for translation or localisation), locale convention (i.e. the extent to which texts are formally compliant with the conventions of the target locale) and verity (i.e. the extent to which the text content is suitable for the target readership and locale) (Lommel et al., 2015). One of the advantages of this quality assessment framework is its flexibility, as it offers users the possibility to customise their metric according to their needs, choosing for instance to conduct a holistic (broad) or an analytic (detailed) assessment of translations, or selecting some parameters rather than others, depending on assessment goals, the time available or other factors (Lommel et al., 2015).

Translation quality standards that are based on industrial quality assurance such as ISO 9000 are attempts to cater for the need for precise, objective, quantifiable and widely accepted translation assessment tools. However, as Williams (2004: xvii) comments, translation is a complex and heterogeneous intellectual product, not a standardised or replicable industrial item, and as a result of this, translation quality assessment continues to be a challenging and controversial area. Among other pending issues, notions such as accuracy, fidelity or mistranslation continue to be debatable; quantification of errors with a view to forming an overall judgement of a translation is questionable; assessors do not agree on what constitutes a major or a minor translation error; and there is no consensus either on whose perspective and interests (translator as an expert, commissioner, end user, etc.) are to be given priority in our understanding and assessment of quality (Drugan, 2013: 38). Within this context, in the following sections a general framework is put forward for quality in community translation, not in the sense of an assessment model or standard, but in the form of general guidelines that would be helpful to follow before, during and after translation for a local community.