Who are you excluding with your default language?

For most open source projects, English is the default language. It’s the language of documentation, chat, email, and the software itself. Content starts in English and is then translated into other languages. (The Fedora Magazine writing guidelines ironically say “US English is the lingua franca for the Fedora Project overall.”) So where does that leave people who aren’t proficient in English?

In defense of a default language

Projects need a default language. If contributors can’t communicate, how can they ever hope to work together? Picking a single language gives you that common basis for communicating. That’s not to say that you can’t have language-specific communities, but the project-level communication happens in the default language.

I’m sure there are projects out there which have a “write in any supported language and we’ll translate” approach. This can work for documentation and the like, but no one is going to sit and translate every email or chat message — or even bug reports — into a dozen different languages. And now you’ve shifted from an “everyone has to know our default language” form of exclusion to an “anyone who wants to contribute needs to know all of our supported languages” form. After all, it’s hard to have a multi-directional translation flow.

Who gets excluded?

I was inspired to write this post after reading an article in Brno Daily about the Czech Republic’s English proficiency. I was surprised to see that Czechs ranked 23rd out of 35. Many of the people I work with on a daily basis are Czech and they have excellent English skills. On the other hand, just over three decades ago Czechs who learned a second language were probably taught Russian, not English. The older population is far less likely to know English and thus won’t be able to participate in most open source projects. (Granted, although it’s toward the bottom of the EU, the Czech Republic was still in the “high proficiency” category).

When you look at the proficiency in other countries, there’s a pretty clear pattern. The low and very low proficiency countries tend to be in the global south, and economically disadvantaged. 60% of G-20 countries had “moderate proficiency” or better. In addition to the geographic differences, there are gender and age gaps.

Picking a default language

The default language does not have to be English. Look at the existing communities that you’re trying to draw from or serve. What language are they most comfortable with? If you’re developing a tool to interact with French government services, for example, using French as the default language seems reasonable. As with most choices, there isn’t an objectively right answer, there’s just the best answer for your project. But bear in mind how the choice you make impacts who can participate. If your language choice reinforces existing biases, what will you do to mitigate that?

I don’t have a magical solution. Picking English as a default language makes sense if you want to attract people who are already contributing to open source projects. But we already know that open source communities lack diversity on many axes. Maybe it’s better to start your project in a different language? Those of us who are privileged enough to be native English speakers must do the work to make sure we’re being inclusive in other ways.

This post’s featured image by Towfiqu barbhuiya on Unsplash.