MT in Linguist’s Eye – An Interview

MT in Linguist’s Eye – An Interview


There are as many opinions on machine translation as there are people. Considering different perspectives and hearing differing ideas is a good way to drive improvement and innovation. That’s why this month, we are interviewing one of TranslateMedia’s freelance linguists to find out what she thinks about machine translation and how MT is shaping translator’s profession today.

Susanne is a native German linguist who has been working with TranslateMedia for the past ten years. Susanne specialises in creative services such as transcreation and is particularly enthusiastic about fashion translations, but also enjoys more technical tasks such as automotive translation. In her spare time, Susanne loves reading anything and everything, listening to heavy metal and playing with her cat Mimi.

TranslateMedia (TM): Thank you for agreeing to share your thoughts on machine translation with us, Susie. I would like to start with your early encounters with MT. Can you remember the first time you came across machine translation? What were your thoughts back then?

Susanne (S): This was quite a while ago but as far as I remember, I tried to read interviews with a Danish musician conducted in Danish. I don’t speak the language so I didn’t understand a word. My passion for music makes me want to know a bit more about the artists I like so I resorted to using Bing Microsoft Translator. It was Danish into English as translation from Danish into German wasn’t available back then. It needed a good amount of intuition and background knowledge to understand what the musician said. For example, I remember that Bing translated the word for “record”, as in music album, as “dinner plate” or “platter”. Perhaps there is a colloquial Danish word that had the same meaning. That sounded rather funny in some contexts.

TM: Has anything changed since that time?

S: I don’t tend to use machine translation extensively in my private life but I do use the online dictionary Linguee for my work and often notice machine translations provided on the website. Some of them are not very good, but some are rather useful for information and research purposes. I can see how MT is getting better but it is not quite there yet, especially in comparison with other smart applications like the “thinking fridge” or Alexa. That is probably due to the fact that language is much more complex and so much less structured than finding a route or writing an electronic shopping list.

Language is complex and often not very structured. For that reason, it presents a challenge for applications such as Alexa.

TM: You mention that some of the machine translations can be useful for information purposes. From your experience, do you find MT more helpful in some instances than others? Is there any particular type of content where it proves most useful to you?

S: If content is repetitive and not too creative MT works well. This applies to well-structured pieces of content such as lists of ingredients, features and descriptions of products, dates, anything not too colloquial. On the other hand, MT starts to struggle with very long sentences, creative wordplays or ambiguous referencing – that is where human’s touch and cultural awareness is indispensable.

TM: As a translator, would you say that machine translation is a passing trend?

S: No, definitely not. It goes hand in hand with the digitalisation of almost every other industry and our private lives and, of course, also with the ever-growing globalisation.

TM: Given that you think machine translation is here to stay, how do you see it affecting the shape of the translator’s profession in the next five years?

S: I can already see a clear increase in enquiries for Post-Editing Machine Translation service. At the very least, agencies are announcing that they are extending and intensifying MT-related services. On the other hand, most end-clients still seem quite wary and prefer human quality over quick and cheap MT. I would say that honest, supportive and open end-client education is very important here as they usually don’t speak the target language(s) themselves. They often need guidance in selecting the most suitable service for their localisation needs.

Wooden signposts with help, support, advice and guidance written on them

Educating and guiding clients in the localisation area is key.

TM: Can you see any everyday applications of machine translation outside of the language industry?

S: Apart from the example I gave about reading interviews with a Danish musician, the other experience with MT that I had outside of my translation work was with eBay’s machine translated product listings. Unfortunately, I can’t say it was an entirely positive one. I was browsing for music records and titles of albums such as AC/DC’s “Highway to Hell” are not translated into “real life” German. eBay’s offering of “Autobahn in die Hölle” certainly made me chuckle. While I can definitely see this feature being useful if you searched eBay pages in countries where you don’t speak the language for other kinds of articles, like clothing or electronics, with proper names such as album titles, MT should be applied with a little more consideration. Keeping the English title of an album alongside a translation would certainly help.

TM: Is there anything else you would like to add?

S: I hope and believe that human translators will never be superfluous and that there will always be a “use” for our kind (of work), even if MT progresses to near perfection. After all, translations by humans most often benefit from revising, proofreading and/or editing as well, if a polished, high-quality target text is required. I do enjoy the creative side of our job the most, and based on my experience, I’m confident that creative human linguists won’t be replaceable either in the near of far future (fingers crossed).

TM: Thank you, Susie. It has been great talking to you and the insight you have provided has been fascinating. Very best of luck in the New Year.

Related posts

Subscribe to our newsletter