Great scientific progress is not the product of the work of a single scientist. Science is built on the summed efforts of thousands of researchers and scholars working in institutions from all around the globe.
For scientific progress to the possible in any area, there has to be a constant exchange between specialists, across borders, and across languages. But how is it possible?
Academic translation is one of many disciplines behind the international exchange of knowledge. But, could the need for translation be eliminated by adopting English as a lingua franca? Could artificial intelligence eventually replace academic translators? In this post, we’ll examine the aforementioned questions from an evidence-based perspective.
English as a Lingua Franca
English is already academia’s lingua franca, much like Latin once was. For instance, in most Western European countries, researchers are far more likely to publish in English than in their native language. But this isn’t good news. And it doesn’t translate into richer communication.
By the way, this is part of our Full Translation Services Guide – if you want to find out more about the different industries where translation services are needed, and the history of this practice and hopes for the future, check it out!
As Karen Bennett from the Center for English Studies at the University of Lisbon, explained in a 2013 article, the dominance of English can prevent certain local scholarly discourses from ever getting the attention they deserve. On the other hand, in order to be accepted for international publication, non-English speaking scholars often have their production “domesticated” to match pre-existing trends and customs in English-speaking academic traditions. This results in what Bennett refers to as “the destruction of the entire epistemological infrastructure of the original”.
When foreign scholars are forced to write in English and conform to the rhetorical patterns of the language, knowledge can be lost in the process. But this isn’t just the case for scholars writing in English. Translation can have exactly the same effect.
Can Machine Translation Do the Trick?
With new development in the field of machine translation and the media attention it attracted, many have started wondering whether machine translation can be used as a quick, low-cost translation solution across many fields, including academia.
The prospect of removing the human element from translations is still very unlikely. As noted in a recent Synced article, state-of-the-art neural networks are capable of translating texts with a 60-90% degree of accuracy. But these results are not at all consistent. For instance, the inputs with which one can train the model for it to make translations can only have the length of a sentence. This leaves us with neural networks that can translate sentences with 60-90% accuracy, but that will fail to create cohesive paragraphs and documents. For instance, a neural network might translate a certain word in three or four different manners throughout a text. This terminological inconsistency can become dangerous when we’re dealing with scientific papers.
A great academic translation might be produced by a neural network that’s doing “the heavy lifting”, but will have its output revised and edited by a highly-trained human. Most machine translation technology used in professional settings nowadays produces “pre-translations”. “Pre-translations” are suggestions for the translator to take or discard. This makes the linguist’s work a little easier – but it’s far from supplanting them.
The Qualities of a Great Academic Translation Specialist
There are many challenges to academic translation. For starters, it’s always best for papers to be in the hands of translators that have worked in the field and are accustomed to the terminology at hand. An academic translator should also have enough preexisting knowledge of the area to be able to deduce ambiguities and come up with the correct equivalent in the target language. Academic translation poses challenges in terms of precision, and of not “eroding the epistemological infrastructure”. Challenges that only a highly-trained translation professional with area expertise can handle.
Academic translators should be familiar with the area of study at hand, and the context of production of the academic paper in question. But, aside from insight into the topic of the document, a translator should also be experienced in translation itself. It’s not enough to know what the paper is about. A translator should be able to properly transmit that information in another language, and that demands a particular skill set. That’s why even for seasoned scholars, a DIY approach is never preferable.
Adopting English as a lingua franca can result in a homogenization of academic output, and therefore, of academic discourse. But even when scholars are producing in their mother tongue, a translation that fails to acknowledge certain nuances can result in a text that’s stripped of certain layers of complexity present in the original.
While machine translation can make certain translation processes more agile, it’s yet to replace a professional translator. When it comes to academic translations, it’s best to rely on an agency for translation services. And if possible, it should be a reputable agency with considerable experience and a solid record working with organizations in your field.