(This post was originally published on the ATA's The Savvy Newcomer blog as a guest post)
Author: Olivia Albrecht
As professional translators and interpreters, we are always striving to provide high-quality services to our clients, be that translation, interpretation, revision work, etc. Yet what does high-quality work look like as a language professional? How can it be measured and how do we know if we are providing quality work? Drs. Geoffrey Koby and Isabel Lacruz tackled this massive subject in their academic introduction to a volume of Linguistica Antverpiensia, New Series: Themes in Translation Studies that focuses on the issue of translation quality.
Their introductory article, The thorny problem of translation and interpreting quality, talks about how translation and interpretation quality is measured around the world with a handful of examples and explains why it is so hard for many professionals to agree on what translation quality really is.
The main problem with discussing translation quality is that there is no set definition nor a widely accepted tool for measuring it. The authors discussed the possibility of two largely acknowledged definitions put forward in an article for Revista Tradumàtica: tecnologies de la traducció:
Narrow definition: “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.”
Broad definition: “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.”
These two differing ideas bring up the question of whether high-quality translation and interpreting is indeed necessary for all projects. Machine translation (MT) and post-editing have made this question even more relevant nowadays. Is it not better to have a translation produced by MT that does not use well-formed language or sound native, but gets the idea across for instances where the text would not have been translated at all? Perhaps, but would that text still be considered quality work? That is where many views differ.
So, despite a lack of a universal basic definition for translation quality, how can one’s translation quality be measured? Different associations and government organizations around the world certify and test translators and interpreters to ensure that they are competent language mediators. However, assessing language professionals varies greatly in form, content, approach, length, etc. for each exam.
Many translation exams are based on either a holistic assessment or a points-off system. The ATA certification exam uses the points-off system where errors of various severity levels have different point values and will be deducted from an overall score. However, Koby and Lacruz state that this system fundamentally emphasizes failure and not what the individual did right. The correct is assumed; the incorrect is pointed out. Yet if full accuracy means zero (or nearly zero) errors, then an argument can be made for preferring error-based assessment over holistic assessment.
In regards to editing and proofreading practices in translation, revisers will often make unnecessary corrections to a translation. This inhibits the accuracy and the quality of the text and also wastes time and money for the client. The authors point out the need for more research in this area that would incorporate explanations from revisers as to why they made changes in order to classify them as “necessary” or “unnecessary” and keep a holistic view of the translations to see how they affect translation quality.
The second half of the introductory article discusses the different articles in the volume, which present ways that translation, revision, MT and post-editing, interlingual live subtitling, and interpreting quality are assessed. For brevity purposes, here are some of the ways that researchers differed in opinion in regards to assessing translation quality alone.
Research from the FBI concludes that there is a third aspect of assessing translation in addition to source language comprehension and target language writing skills. Translators that produce quality work also possess translation proficiency, a separate ability to translate well, which must also be assessed.
Another set of researchers believe that translation quality can be determined by looking primarily at the target text, as opposed to measuring the adequacy of the transfer between languages. They assessed this through the use of corpora and extracted several features to be analyzed. The researchers concluded that this method, in addition to constructive feedback, would be a better approach to assessing quality in translation.
Yet two other researchers disagree with both of these theories and suggest that a Calibration of Dichotomous Items (CDI) method is more appropriate for assessing translations. This method takes translations of the same material from a large group of translators and identifies the segments where there was a large difference in translation quality. Then, they decide which translations are acceptable and which are not, but they do not attempt to rate the quality of the translations in a more refined way.
A final set of researchers analyzed the different testing approaches for translators in Finland against the testing systems in Sweden, Norway, and the German state of Bavaria. After assessing the different approaches to testing in these other countries, some of which use error analysis method and others a criterion-based method, the authors decided to improve the Finnish examinations further by proposing a simplified scoring chart.
Though it is unclear which methods of assessment are the most accurate, this introductory article and the other articles in the volume were meant to shed light on some of the various ways that translation quality can be tested and the reason why it is so hard to define quality in language translation. Human language and mediation are complex, therefore quality assessment for translation, interpreting, and related activities remains a thorny problem.