skip to Main Content

Post-editing and the changing role of the skilled translator

Post-editing, or the editing done to improve machine-translated content to a publishable quality, has long been part of the translation repertoire in one form or another. However, with an increasing presence of machine translation (MT) in our everyday lives, there has been recent debate and uncertainty about the role of the translator vis-à-vis MT and post-editing.

The 90s presented a similar scenario with the introduction of translation memories (TM) – tools that store and translatorsuggest previously translated content so that the translator doesn’t have to translate the same text over and over again. While there was some resistance to TMs initially, others were keen to adopt this new tool into their toolbox and reap the benefits, invest in training, and point out what TMs could and could not do.

This now commonplace tool has brought with it gains in productivity, more efficient resource management, and incredible value in research and development of MT itself – popular data-driven methods like Google Translate are largely reliant on human translations! Yet, there were (are) fears of blindly following TM suggestions and of the damage such tools would do in reducing the value of the translator in the translation process.

Two decades later, the push for post-editing with MT presents a similar argument for many: adopt a new technology that may improve productivity but risk minimising the role of the translator? Promises of productivity gains may come hand-in-hand with the risk of being paid less for a perceivably ‘easier’ job. There are arguments that translators may be paid X% less to post-edit, but will work Y% faster and earn the same or more than usual, but substantiation for this remains to be seen. On the other hand, the translator may even spend more time correcting repetitive errors from MT than ordinarily necessary, or retranslating what a system has done incorrectly.

Such a question is being asked in many other professions – at what point is there a shift from the machine assisting the worker to the worker assisting the machine? In other words, a question of agency in the natural reluctance to give up some control in the face of uncertainty for an envisaged return on investment that may not be of equal value for all.

Remedies for uncertainty are education, exploration, and even trial and error. In this sense, by both being allowed more interaction and themselves being open to such interaction, translators can explore what works and doesn’t work in adopting MT. From an organisational point of view, it is vital that rather than feeling forced to adopt a tool, one has the ability to truly feed into and influence a process – this is immensely rewarding for everyone involved.

Recent European Commission funded collaborative projects such as QTLaunchPad work closely with translators and language service providers to incorporate their views, experiences, and usage of translation and language technologies including MT and post-editing, so that we can all focus our efforts on overcoming the challenges of current technologies. Industry-academia collaborations are also excellent ways to strengthen these connections and keep communities in open dialogue with one another, for instance, by empowering student translators to explore what cloud-based MT can and cannot do for them as they enter the translation workplace.

At the core of these approaches is the skilled translator, who is adding value to the process with a thorough understanding of the strengths and limitations of these tools and technologies. This is coupled with the inclusion and ownership that allows for open and worthwhile collaborations. In what is becoming an increasingly technology-enhanced (albeit technology-reliant) world of work, it is vital for any profession to be fully equipped to understand, use (or not), and even improve upon tools that can add value to its work.

Further Reading:

●  Doherty, S. & Kenny, D. & Way, A. 2012. Taking statistical machine translation to the student translator. Proceedings of the 10th Biennial Conference of the Association for Machine Translation in the Americas. San Diego, CA.

●  Doherty, S. & Moorkens, J. 2013. Investigating the experience of translation technology labs: pedagogical implications. Journal of Specialised Translation, 19, pp. 122-136.

●  Kenny, D. & Way, A. 2001. Teaching Machine Translation and Translation Technology: a Contrastive Study. Proceedings of the MT Summit VIII Workshop on Teaching Machine Translation IAMT/EAMT, Santiago de Compostela, Spain, pp. 13-17.

●  Lagoudaki, E. 2008. The value of machine translation for the professional translator. Proceedings of the 8th Biennial Conference of the Association for Machine Translation in the Americas. Waikiki, Hawaii, pp. 262-269.


Stephen Doherty

This is a guest post by Dr. Stephen Doherty, a post-doctoral researcher at the Centre for Next Generation Localisation in Dublin City University. He researches and lectures in areas of language and translation technology, cognition, and human-computer interaction. He is currently working on a collaborative EC-funded research project QTLaunchPad, and teaching translation technologies with a focus on MT.

Back To Top