Facilitating multilingual crisis communication: the applications of speech-enabled machine translation post-editing (PEMT) in crisis translation

Doctoral project - Claudia Wiesinger, MA MA

In sudden-onset disasters, effective communication with affected communities often requires translation and interpreting (T&I). However, the need for multilingual communication sometimes cannot be met as a result of insufficient communication policies, a dearth of language professionals and/or multilingual first responders, time pressures, the disruption caused by disaster events, or a combination of the above. This can have dire consequences for those who do not speak the local language or the lingua franca.

And yet, language service provision is not just problematic in the early stages of sudden-onset disasters. When disaster events – natural, human-caused or otherwise – are not contained, they can become cascading crises with negative, cascading effects that are felt long after the initial disaster event itself. Here, too, language barriers often prevent timely and effective communication with vulnerable populations

One of the main challenges in crisis translation relates to a central dilemma: the need for both quality and productivity. In disasters and crises, it is paramount that highly sensitive content be translated accurately. However, in high-pressure situations, the sheer volume of texts and frequency of updates can adversely affect quality, and the lack of language professionals poses an additional challenge.

In light of the challenges that arise in different crisis phases and the current use of language technologies, this doctoral project explores the vital but often overlooked role of technology-supported translation practices in cascading crises with the aim of facilitating the work of professional and citizen translators. More specifically, it investigates the integration of speech synthesis into existing crisis translation workflows that make use of machine translation post-editing (PEMT). The underlying research hypothesis is that the use of speech synthesis will have a positive effect on the quality of the post-edited output and the translators’ productivity. The findings of this research will form the basis of guidelines for professional practice.

The project will use a mixed-methods approach, i.e., semi-structured interviews, questionnaires and an experimental study, to:

  • identify current challenges in crisis translation workflows that the use of speech synthesis can address;
  • test a novel crisis translation workflow using PEMT and speech synthesis;
  • develop guidelines based on best practice for the effective and appropriate use of speech-enabled PEMT in crisis translation.