#L10n

This cryptic title is the abbreviation used for the activity I’ve been busy with for the past months: localisation (L + 10 letters (count them!) + n — clever, innit?) But what is localisation? Mainly, it is the name given to software, videogame and website translation. Why a different name? Because although localisation is a form of translation, with all the cultural adaptation it entails, it has technical specificities which I will discuss briefly.

Logo des Warlocs
https://warlocs.com/

First, a bit of context: last January, Warlocs, an independent collective of translators, were kind enough to entrust me with the translation of Not Tonight 2. In this game, you play three friends travelling across dystopian United States to save a fourth one who was captured by a group of quasi-Trumpian extremists. To fund their journey, they work as bouncers: the core gameplay is time-limited document verification, inspired by the most excellent Papers, Please.

So I enthusiastically dived in the pages… or rather the files. As usual in this field, some 70000 words were sent to me as spreadsheets, where each line can represent a bit of the interface, a narration paragraph, a dialogue line, etc. It is a bit daunting at first, but you get used to it. During the translation, I encountered three main difficulties that were specific to localisation.

Feuille Excel de localisation
So I heard you like words?

First, the character limits. The expansion rate, well-known to book publishers, must be watched very closely when localising. For instance, a book in English usually grows by 10 to 20% when translated into French. Although it is acceptable for dialogue lines, it can be absolutely impossible for UI elements which have only limited space (see for example how many letters there are in the French translations of “run”). That’s why in the picture above, there are three columns on the right side which I use to count the number of characters in English, in French, and the expansion rate, which turns red when it goes above 10%. I was always watching it when translating.

Next, the problems with spreadsheets themselves. Translating a text is usually quite linear, but it’s not necessarily the case here: the authors’ texts, often written with other tools, may appear in an arbitrary order once converted to spreadsheet format. Some dialogues must then be translated out of order, which makes any form of consistency and flow hard to achieve. Another issue: the identifiers of each line (Who is speaking? In what scene?) can lack clarity or contain errors, which means you have to be extra watchful. Thankfully, the developers had given us a beta version of the game to help us evaluate the context of each line, but it is a luxury that most localisers are not given.

Finally, you can’t always translate everything. Although the spreadsheets contain most of the game’s texts, it’s sometimes not the case for words appearing in background pictures. Let’s take an example from Not Tonight 2: there is a cult called “the Creed cult”, which you could translate by “la secte du crédo” (“crédo” is “creed” in French). However, the word “Creed” was present in the level background, and not translatable, so I’ve had to keep it and work around it. My translation then became “la secte Creed”, while its members, the “Creedsmen” in the original, became… “les Creediens”. I like this translation because it has religious overtones — “Creediens” sounds like Chrétien (=Christian) in French — but I might not have used it if I could have translated everything.

There are ways to prepare all the game’s elements for them to be translatable (this process is known as internationalisation, or i18n — i + 18 letters + n — can’t get enough of this). It is not always possible, however, or even desirable, and it can represent an unwelcome cost for a small studio.

Meme de bob l'éponge fan de tableurs
Me after several weeks’ work.

Localisation implies quite a few more constraints: one among these (which I didn’t have to deal with in this case) is translating lines that will be recorded by actors, which entails all sorts of problems that audiovisual translators know well. In the end, although all these parameters sometimes feel like you were given an impossible mission, localisation is a fascinating job, and one I really love. I will have more occasions to write about it, and to dive deeper, because I will work with the Warlocs on several more projects. In any case, if these few words allowed you to have a quick look behind the scenes, be sure to be indulgent next time a localisation seems a bit off: the translator has surely done everything they could!

Hallowed be Their name

On this blog, I would like to write about the translations I’m working on, both to show how the sausage is made and to bring myself to explain my translation decisions. This way, I’ll be able to further convince myself — and maybe you too — that they are relevant. (At a given time, of course: you always find fault with a translation you come back to later on.)

I’ll start then with the translation I’ve worked on for my Master’s degree: The Breath of the Sun (Aqueduct Press, 2018), by Isaac Fellman. There were several difficult aspects to this book, and I wasn’t sure which part to present here. The short-lived (and dishonest) crusade against the French neutral pronoun “iel” (which you can read about here) helped me decide, since I did use the infamous pronoun in my translation.

First, some context: the narrator, Lamat, has recently met Disaine, a scientist and former priestess. Disaine wants to hire Lamat as a guide to climb the mountain sitting atop of their world, the top of which, legend has it, goes beyond the stratosphere. Lamat belongs to the Holoh people: this mountain is a god to them, and they can’t climb it without respecting some complex ceremonies. Disaine implies that Lamat hasn’t had much trouble breaking her people’s taboos.

Here is her answer:


The original (p. 14):

“It wasn’t that,” I said, and looked over the basket at the side of the mountain, with its billion footholds in the snow. Snow on God’s body, dry and fine. “It wasn’t easy to break at all. But I thought — and I still think, even though it was such a disaster, even though people died and marriages ended…”
“Yes?”
“I grew up being told that God doesn’t want us to climb. That we wound Them with our feet, that we blood Them with our fingernails. And that I’m not sure it’s true. The Holoh are the only people who are visible to God. Why would They choose us, if not so that we could someday see Them face to face?”

My translation:

– Ce n’est pas ça », ai-je répondu en regardant, derrière la nacelle, la montagne et ses milliards de points d’appui dans la neige. De la neige sur le corps de Dieu, sèche et fine. « Ça n’a pas été facile du tout. Mais je me suis dit – et je me dis encore, malgré le désastre, même s’il y a eu des morts et des mariages brisés…
– Oui ?
– On m’a toujours dit que Dieu ne voulait pas qu’on grimpe. Que nos pieds Læ blessaient, que nos ongles L’écorchaient. Et je ne suis pas sûre que ce soit vrai. Les Holohs sont le seul peuple visible aux yeux de Dieu. Pourquoi nous choisirait-Iel, sinon pour que nous puissions un jour Læ voir en face ? »


Lamat uses the singular “They” to talk about God/the mountain, and the capital T reinforces the divine character of that being. Her using it is all the more interesting since Disaine uses “He”, the male singular pronoun, to talk about that same god. Theological discussions are frequent in the book, and it is important to render them in all their nuances. There is thus no way I could neutralise this difference in French and have both characters use the masculine singular pronoun il, all the more since gender is a central theme of the book.

At this stage, the translator worries. How are you supposed to translate this, in a language supposedly so hostile to gender-neutral terms? If using the singular they is perfectly appropriate in English (since at least 1375, the Oxford English Dictionary tells us), French is sorely lacking in the neutral pronoun department… at least that’s what I thought. Further research in the matter brought me to iel, of course, but I’ve also encountered ael, al, el, ol, ul, yel and even ille. Some of these are regionalisms, some already existing in Middle or Old French while some others are more recent creations, but there are many possibilities.

Close-up on "they" from William & the Werewolf (1375)
The use of the singular they is so old that it was first written þei ! Yes, it is actually older than the famous th which has caused so many pains to French students.

However, these are pronouns that few French readers are used to encountering. My translation would then automatically feel unfamiliar in a way that the original doesn’t, which is why I decided to use “iel” so as to make that discomfort minimal: a French reader understands it is a mix of il and elle, and it sounds common enough in French to not sound too alien.

That being settled, I still had to deal with “Them”. Again, the translator worries. It would indeed be regrettable to use the gender-neutral pronoun iel only to then pick the gendered objet pronouns le or la which would bring back Lamat’s neutral god to a male or female gender. More research led to more discoveries: I had at my disposal object pronouns such as Lo, Lea, Lu or even Lae. Lamat’s people feeling tradition-oriented and speaking a language that is somewhat archaic, I opted for using Lae and adding a ligature to get (rather rare in French, tied letters mainly appear in borrowings from Latin, which gives this pronoun a more historical feel).

This aren’t necessarily pronouns I use on a daily basis. I seldom read them, and hear them even less frequently, but I felt they would be the most appropriate option for this text. Of course, a reader who has never come across them will probably stumble at first: they will wonder briefly what iel is referring to, how to pronounce… But they will also perceive another aspect of what differentiates the two main characters, and it is my hope that they will understand better the book and its world because of this decision.