One of the things I sometimes see being ignored in video game translation is the existence of gendered languages. This matters a lot because, depending on the language, it can make or break what would’ve been an easy job. These issues are now less common than they were a few years ago, but they still happen from time to time, so it should be worth it to talk a bit about them.

---

Genders in grammar

So, here’s the thing: English is a non-gendered language. Despite having pronouns, which are gendered (obviously), nouns and adjectives don’t carry any quality of grammatical gender. A city is a city, a dog is a dog, a cat is a cat, a car is a car, a person is a person: all of these words don’t have any “inherent” gender, they aren’t associated “inherently” with any pronoun until they are given a specific form, and even then they can just carry an “it” or a non-plural “they”.

That isn’t the case in Portuguese (as with many other languages): city is a feminine noun (a cidade), dog can be a feminine or masculine noun (o cachorro / a cachorra ), cat too (o gato / a gata), car is a masculine noun (o carro), and person is a feminine noun (a pessoa).

That last case might throw a few people off because, after all, what if the person is a man? Well, it doesn’t matter, the word is gendered, not the concept it represents. Grammatical gender is just a way to classify words, it doesn’t necessarily mean the gender of what it represents. In the case of dog and cat, it also denotes their biological gender, but a city isn’t female, the same way a car isn’t male. Again, it’s just a way to classify words.

Why does grammatical gender matter then? Because it dictates the grammatical concordance of the whole sentence: the noun’s gender dictates which article to use, and how to decline the adjective so it can be properly related to the noun. And Portuguese has it easy: in a few other languages, gender dictates declension of every word in a sentence, and sometimes even the verb (some languages have gendered verbs, which is nuts, but also makes sense).

Now, when something is referring to a specific person, for example, grammatical gender also tells us the gender of that person. “The beautiful man” is “o homem bonito“, “the beautiful woman” is “a mulher bonita“. To say “o homem bonita” is grammatically wrong and doesn’t make it clear if the person is a man or a woman. On the other hand, “the beautiful person”, “a pessoa bonita“, can be used to describe either a man or a woman — despite the sentence being grammatically feminine — because the word for “person” (pessoa) is a “generic concept” and not linked to any human gender.

In the example above, the word “bonito” or “bonita” is an adjective, so it must follow the noun’s gender. In English, this isn’t an issue because, as stated at the beginning, English isn’t a grammatically gendered language, so “beautiful” works for anyone.

I won’t talk about pronouns because their use is well understood by anyone who speaks English (and can read this post). It’s worth mentioning, though, that the neutral pronouns “they/them” don’t work for every language, which is another layer of complexity: in Portuguese, for example, plural pronouns are also gendered, so they can’t be transposed to a singular gender-neutral like “they”.

---

The issues

Ok, so, why does a language having grammatical gender matter? Because, since adjectives are gendered, every speaker in a dialogue must be identifiable and gendered so the sentence makes sense. This is something that most people are aware of, but not to a deeper degree to consider multiple improbable scenarios that could very well happen. For example, most people are aware enough to signal in a dialogue when it’s a man talking to a woman, or a woman talking to a woman, and so on. But then, there’s a man talking to a group. Is this a group of men or women? If it’s a mixed group, are there more women than men, or vice-versa? It might seem silly, but this matters when trying to have a clean and well-written line that needs to follow grammatical gender rules.

Now imagine that the other characters are talking with the protagonist: is the protagonist a man, a woman, or non-binary? This could be important when describing actions the protagonist does (remember, some languages have gendered verbs!) or talking about some of the protagonist’s traits (remember, some languages have gendered adjectives and articles!).

Let’s take it one step further: how does the game communicate with the player? It’s easy, in English, to keep every menu item, action, description, and so on, in a neutral gender, but that won’t work for every language. So, will the game ask the player for their gender? If not, is the player’s gender important? In Portuguese, the masculine gender is the “default” because it merged with the neutral gender when it “evolved” from Latin (the neutral and masculine genders from Latin melded into the modern masculine gender for Portuguese), but since it’s associated with a male presence, it could be offensive or disrespectful for non-male players.

Sadly, there’s no easy way to deal with this issue. Ideally, every line of dialogue that involves someone should be duplicated in case this someone can be of any other gender. Every character should be tagged so that during the interaction, the line with the right gender is called.

It should be possible to use variables just to change relevant gendered words, calling them from a pre-translated table, but this solution could quickly turn into a combinatory puzzle that’s hard to predict since we don’t know every language in the world. In a perfect world, everything could’ve been checked on a case-by-case basis.

Well, we are talking about characters, but what about items? Remember, in many languages, every noun is gendered, including objects, so if the game has items with names comprised of nouns and adjectives (like an Awesome Fire Sword of Joy +10), what should be done? If the game has a procedural naming system, how should it be set up so the gendered adjectives don’t mix with the wrong gendered nouns?

With items, things are somewhat simpler when dealing specifically with UI, like inventory: if the item names are set and don’t change, they can be translated whole as text, so it’s translation as usual. Now, if their names are procedural, there should be a list of variables for nouns (like sword, axe, helmet, etc) and another one for adjectives (fire, heavenly, ghoul-killing, etc). The list of adjectives should be duplicated, so they can match with nouns from any gender. Also, it should be possible to rearrange the variables, so the translators can set the noun and adjective in the right order for their language — in Portuguese, for example, adjectives come after nouns, while in English it’s the opposite.

It’s worth noting that, if the item names appear in dialogues and quests, there should also be a variable for the article that precedes the item, as many languages also have gendered articles. Needless to say, they should have their own translated table to be called from. Some languages, though, don’t use articles at all, so the translator should have the option to add or remove the article.

---

Conclusion

As it’s possible to see, there are a lot of things to consider that, if left as an afterthought, will be nigh-impossible to implement retroactively in a timely and orderly manner. This is, obviously, not extensive and relates mostly to my experience translating from English into Brazilian Portuguese.

Also, I can’t talk specifically about the technical aspects of importing and exporting text or dealing with variables, because that changes depending on how the game is being built and what tools the engine supports. Suffice it to say, though, that hard-coding the script into the code is a bad idea: the text should be called into the game from a table, XML file, or something similar, that can be easily edited if needed. Most games I translate, some with millions of words, have their script set in a .xls, and each line has its own ID and some identifiable markers, like character and gender.

It should also be noted that most of these issues are present in games with branching storylines and player-created protagonists. If a game has a set story that will never change (like kinetic visual novels), or doesn’t have an extensive inventory system (like shmups, some adventures, and so on) some issues mentioned are pretty much non-existent.

Once, I worked on a management mobile game where the engine wasn’t built to handle multiple genders, so I had to butcher the Portuguese grammar (way before the current rule-breaking was acceptable) to make everything “neutral”, and UI lines (like interactions) had a very inconsistent format. It wasn’t fun.

On the other hand, some colleagues translated a game a few years ago where every character was non-binary, so they used the ruleset for the elu system, one of the acceptable rule-breaking systems. In another case, the player and every other character could be male, female, or non-binary, so every line was tripled.

Recently, I worked in a few games where some barks weren’t gender-tagged, so they had to be translated in a neutral way. In this case, the team and I were able to “walk around” the Portuguese limitations avoiding pronouns, and only using adjectives that didn’t vary by gender [usually, words ending in -e in Portuguese can be either male or female (or neutral)].

It’s impossible to predict every issue that could arise from the text, especially on more complex scripts, but at least being informed of what could happen goes a long way to guarantee a smoother translation.

---

PS: I focused on English as a non-gendered language in this post because that’s the “default” language for video game production nowadays, but linguistically, English isn’t that much of an outlier. Also, it just so happens that a lot of Western-language speakers who buy video games speak gendered languages natively, conflicting with the English production.