The Struggle with Translating Linguistic Variations

   

It is one of the biggest challenges for a translator to overcome. You’re reading a text for the first time and all of a sudden, a character starts talking with a strong Scottish accent and you don’t know what to do. Or sometimes, the entire text you’ve been sent is written in a specific dialect and you can’t see any equivalent in your own language. Most translators choose to solve these issues by using the standardized version of their language. But isn’t something lost in the process somehow? Language surely conveys meaning, but does that mean it's optional to translate the meaning conveyed by linguistic variations?

Dialects

 “If you talk to a man in a language he understands, that goes to his head. If you talk to him in his language that goes to his heart.” (Nelson Mandela)

Many companies believe that focusing on the language they use can make a big difference in the way they are perceived by their customers. Some are even starting to use localized translation services in order to communicate in a way that goes directly to the heart of any customers passing by.

Dialect does not only mean a regional speech pattern. It can also be defined by other criteria, such as age, gender, social class or ethnicity. For example, there are over 50 different spoken dialects in China, and most of them are mutually unintelligible from each other. While many businesses translate their website into Chinese (well actually, it’s in Mandarin), a significant number of Chinese people still wouldn’t understand it easily.

Accents

From Harry Potter to The Adventures of Huckleberry Finn, literature contains countless examples of characters with accents which make us glad we didn’t have to translate it ourselves. On the one hand, using the standardized language can be a tough choice when it implies cutting out half of the character’s identity and texture. Some languages such as French have very few accent variations, with the exception perhaps of the southern accent or rural patois. On the other hand, try to transcribe a foreigner’s accent in a French book, for example a French person of African origin, or a character with a Spanish accent: it can easily turn out to be, at best, unreadable, or at worst, racist. Quite often, the translator will choose to write simply something like “He spoke with a Spanish accent”.

In other cases, some accents are identified in the source language, but have no equivalent in the translated one, such as in Emily Bronte’s Wuthering Heights. Joseph and his heavy Yorkshire drawl can hardly be translated and it requires creativity not to lose what was added to his character by his accent in the original text.

If there is no equivalent in the target language, then the real matter here should be: is the accent a defining detail to understand the character? What would be lost if he spoke normally, and can it be shown through other details, such as a physical description or character development? When the variations cannot be translated directly, there are other ways to convey the original meaning when it gets too tricky.

“It just can’t be translated!”

As if it wasn't already hard enough, in some languages, there are other variations than dialect and accent. For example, in Japanese, there are more than 4 different ways to say “I”. The most commonly used being:   

- “Ore”, the most manly one

- “Boku”, the one used by young boys

- “Watashi”, the neutral but a little bit feminine one

- “Atashi”, the truly feminine one     

In the award-winning animation movie Your Name by Shinkai Makoto (2016), a high school girl begins switching bodies intermittently with a young boy when they wake up in the morning.

During one of the first scenes, the girl while inside the boy’s body, is surrounded by his friends and starts a sentence with “Atashi” (the very feminine “I”!) which immediately triggers the surprise of the boys. She understands what’s wrong and tries all 4 versions in order to find the boy’s usual way of speaking. Translating this scene was certainly quite challenging in many languages. While the subtitle makers tried to explain the difference between the 4 types on the go, between brackets, the dubbing team had to find another way. So instead, the translators from different countries came up with all sorts of solutions. The English dubbing slightly modifies  the conversation, and goes like this: “A girl like me…” / “A girl-” / “I mean-” / “A guy like me!”, but some others chose for instance to have the girl’s voice actress to play the boy’s voice as well, at least for the scenes when they are switched. For that scene in particular, she speaks in a very feminine voice at first, and then gradually reduces the feminine aspect in her intonation. All that trouble just for one linguistic variation that could be neither translated, nor ignored. This reminds us that the most important thing is to translate the content even if the style can’t be kept the same.

Could all these tricky variations be a flaw in the system, thus allowing us to take advantage over automatic translators? If human translators bang their heads together when dealing with variation, accents and dialects, it goes without saying that machines are not yet able to do it. Could it be considered some wiggle room thanks to which humans are still guaranteed to have the upper hand on AI?

Sources:

http://dialectblog.com/2011/05/18/dialect-transcription/

https://mappingignorance.org/2014/02/03/mandela-was-right-the-foreign-language-effect/

https://asianabsolute.co.uk/blog/2018/04/24/languages-spoken-in-china/

https://www.fastcompany.com/40585591/how-language-shapes-our-perception-of-reality

https://www.bubblestranslation.com/the-value-of-dialect-in-translation/

Diane D