Translation studies at a university should prepare future translators for work. However, some scholars, including Kristiina Abdallah from University of Eastern Finland, remark that traditional methods for quantifying quality of translation works widely used in universities, such as method of Jualiane House or Christiane Nord, are not always useful in the working life. The reason for this is that requirements for translators and translations alike have significantly changed over the past decades. On the one hand, tight deadlines force you to work faster, but on the other hand, new translation technology should make the work easier. This raises the question: where is the translation industry actually going?
There is an ongoing debate about the possibility that machines will replace humans, but can this be true in linguistics? It might sound surprising, but even Google Translate is getting very accurate (The Washington Post) and can be a good option for human translation, especially when it comes to costs. Language still has a lot of ambiguity, and for example poetry might cause problems for automated translation tools, but after phrase-based statistical machine translation was replaced by neural machine translation, the overall quality of machine translations has increased significantly.
For those who may be interested, here are some machine translation tools recommended by computer scientist and researcher Philipp Koehn:
- Nematus (based on Theano)
- Marian (a C++ re-implementation of Nematus):
- OpenNMT (based on Torch/pyTorch)
- xnmt (based on DyNet)
- Sockeye (based on MXNet)
- T2T (based on Tensorflow)
I think a claim that the original role of a translator will change in the near future is justified in a situation of rapidly evolving technology. This is however something that might not surprise people who work in translation. Finnish researchers Koskinen & Ruokonen state that “in the past three decades, translation has rapidly shifted from a predominantly humanist profession to an increasingly technology-driven practice”. Maybe some day applying translation quality assessment methods, such as mentioned in the beginning of this article, will be the main part of translators’ work.