Subtitle Translation in French: Wordplay and Machine Translation
If you are learning French, chances are you are watching French films with subtitles. Perhaps French cinema is what motivated you to study French in the first place. Even the novels of a single author can motivate students to study an entire language, as seen by a Japanese teacher I met, who said that some of her students had decided to learn Japanese because they liked the novels of the writer Haruki Murakami. French cinema has definitely been influential around the world and directors from many cultures watch it for inspiration.
Subtitles, or sous-titres, certainly make films accessible to a wider audience and are a great help with learning languages. They also present certain issues, being limited by the time it takes for certain phrases to be uttered in the source language and the amount of space available at the bottom of the screen. Many people note that French translations from English appear to be five times longer than the source language. While this might be a slight exaggeration, it is true that the word streaming, for example, would be diffusion en continu, at least in Canada. (The French often use the English word streaming.)
Subtitles in Operas
As indicated by its prefix, the word subtitle refers to written words that appear at the bottom of an image projected onto a screen. These words render into written language the dialogue that takes place in the film. There also exist supertitles, which are used at venues such as opera houses. These are projected onto a panel above the stage and translate what is being sung. Ergo, if you are watching George Bizet’s Carmen at the War Memorial Opera House in San Francisco, you might see “Love is a rebellious bird” on the supertitle panel while hearing “L’amour est un oiseau rebelle”. The discrepancy between the length of the words in English and French is negligible in this case, but even if there were more of a discrepancy, this still would not pose that great of a problem at the opera house, as an opera’s libretto will often have repeated words, so even if the length of words in one language differs greatly from the other, there is still time to display a good translation. Also, most operas do not have very complex plots, often involving people falling in love, facing obstacles, and either ending up married or dead by the end of the final act. This lack of complexity makes a translation less vulnerable to being cumbersome.
Subtitles in Silver Screen
When it comes to the silver screen, subtitles can pose more of a challenge. A film could have a lot of dialogue and a complex plot, making the translators responsible for conveying in the target language, with limited amounts of time and space, utterances and their underlying meanings and any cultural relevance that is evoked in the original. Furthermore, translating Kelly Asbury’s 2017 Smurfs: The Lost Village is very different from translating Wes Anderson’s 2014 The Grand Budapest Hotel, as Bérengère Viennot notes in a 2018 Slate article on the inadequacies of Netflix translations. It is interesting to see how translators have dealt with wordplay in French, and two films, Le Dîner de cons and Amélie, reveal ways in which underlying meanings in words and phrases can be rendered into English.
How Translators Deal with Wordplay
Now, let's take a look of how translators deal with wordplay.
1) Le Dîner de cons
In Francis Veber’s 1998 film, Le Dîner de cons, there is a scene in which a character who is not too bright accidentally telephones a woman he has not meant to contact. The woman at the other end of the line says her name is “Marlène Sassœur”, but he misinterprets this as “Marlène sa sœur”, and thinks that she is the sister of the man whose telephone he is using. He subsequently gives her sensitive information that is part of a sequence of quiproquo in the film. In this case, to keep consistent with the plot, the translators chose to give the woman the name “Marlène Hissister”, which could then be misinterpreted as “Marlène his sister”. I find the translation effective, and would only say that Sassœur in French sounds more like a “normal” name (there is a famous Swiss linguist named Ferdinand de Saussure), whereas Hissister doesn’t really sound like an English name you would come across. Of course, at the San Francisco Opera, you could easily come across someone named Rainbow Bunnyfur, so it’s all a matter of perspective.
2) Amélie
The translation of a scene in Jean-Pierre Jeunet’s 2001 film Amélie illustrates what could potentially be a great puzzle for translators, that is, how to convey the peculiarities of the French language through an English medium. When Amélie’s mother is teaching her to read, Amélie reads aloud the sentence: “Les poules couvent souvent au couvent”, but mispronounces the adverb and noun “souvent” and “couvent”, treating the -ent endings as silent, as they are in the verb “couvent”, which gives us /le pul kuv suv o kuv/ rather than /le pul kuv su vɑ̃ to ku vɑ̃/. The translators interpret Amélie’s version as “Four hens brood hens forth” rather than “The hens brood here henceforth”. It works pretty well, except that “The” changes to “Four” in the section Amélie reads before her mother says “Très bien”, while her reaction to the rest of the sentence, which only contains a minor phonetic difference from the “correct” version, as well as an elimination of one adverb, is to whack the table with a ruler.
The Problems of Machine Translation
1) Word-for-Word Translation
I am assuming that the translations for the above films were completed by trained humans, but this is not always the case for every on-screen translation you may come across. Viennot’s Slate article describes the way in which trained professionals may be overlooked by companies who are trying to produce translations for an enormous output of material at a low cost. In addition, certain sites such as YouTube use speech recognition technology to automatically generate translations, which can become problematic for several reasons. One is the amount of dialogue that flashes onscreen in a very limited amount of time, making the translations hard to read. Another is the word-for-word translation that may become garbled text if the dialogue consists of someone interrupting himself or changing trains of thought mid-sentence. When the Canadian Prime Minister Justin Trudeau said to a woman during a 2018 town hall meeting in Edmonton, Alberta: “I wish you and your organization all the best”, the onscreen French translation on YouTube was: “je vous et votre souhait organisation tout le meilleur”, which doesn’t make a whole lot of sense. A human would probably have translated this more along the lines of “je vous souhaite la meilleure des chances, à vous et à votre organisme”.
2) Text that Comes out of Certain Context
Another thing that escapes automatically generated translations are the humorous situations that come out of certain contexts, such as town halls. During Trudeau’s exchange with the woman he wished all the best, he attempted to mansplain that it was preferable to use the term “peoplekind” rather than “mankind”, as she had originally stated. The French subtitles diplomatically translated these terms as “humanité” and “humanité inclusive”. While this did not convey the reason for the titters in the audience, it did show the automatic translator’s capacity as a moderating force. Perhaps automatically generated subtitles will continue to moderate the great divides between the people who constantly use them. For the moment, we will have to find a French version of the term “mansplain” that is under five words without simply adopting the English term – in Canada, that will just not do.
You May Also Like: