Post-Editing in the Wild: An Empirical Study of Chinese-to-English Professional Translators in New Zealand

Abstract

Globally, increasing adoption of machine translation in the language industry has seen post-editing develop into a mainstream service. In New Zealand it is an emerging practice, likely to grow due to heightened awareness of the need for reliable and competitive language services to serve diverse needs of immigrant communities and business demands. Locally, though, Translation Studies is a maturing discipline with limited availability of formal, research-informed, translator training. Against this backdrop, a small-scale in-depth empirical study on post-editing was conducted, using a frequently machine-translated yet not commonly researched text type. It seeks to shed light on how, and to what effect, local Chinese-to-English professional translators post-edit, in naturalistic conditions, a Chinese patent text machine translated into English. Questionnaire and direct observation methods elicit from eight participants process and product-related data for analyses, employing mainly descriptive statistics. The impact participants' professional attributes, attitude towards machine translation, and dominant language, have on post-editing performance is also studied. In line with prior research, considerable variation is observed among participants in terms of post-editing style, productivity and quality. While participants declare no previous post-editing experience, the study somewhat surprisingly captures evidence of frequent and conspicuous use of free online machine translation among them. Results suggest that participants with a more positive attitude towards machine translation tend to spend more effort improving its output and achieving better quality. Participants who display higher productivity tend to have in-domain translation experience as well as formal translator training. However, there is no evidence of first-language translators with English as their dominant language performing any differently from second-language translators in terms of post-editing productivity or quality. This is a participant-centred exploratory study with known limitations, and is a first New Zealand-based investigation of its kind conducted in authentic settings with local professional translators. Nevertheless, it provides an empirical basis towards developing strategies for future post-editor training, including L2 translators. It makes a timely contribution to the body of knowledge for advancing the understanding of post-editing as a growing practice and as an increasingly important form of human-machine cooperation.