Building the future of translation, online

Many of the conversations at the Berkman Center orbit the topic of technological determinism. While no one will actually admit to being a technological determinist, it’s pretty common to hear arguments that a new technology will lead, inexorably, to a change in human behavior. For instance, we’re forced to redefine our ideas of privacy in an era of social networks which encourage/demand revelations about our personal networks. (Interesting piece in today’s Boston Globe suggests that researchers at MIT are able to make pretty good guesses about people’s sexual orientation based on their network of friends on Facebook and knowledge about the sexuality of some members of the network.) Some of my colleagues argue that we’re bound for a future where potentially embarrasing personal information is so pervasive that we’ll need to work out new social norms about what’s acceptable to remember and what we’re socially required to forget.

By spending the last decade or so watching technology adopted, rejected, refashioned and hacked in the developing world, I’ve developed the strong sense that nothing is inevitable simply because it’s possible, and an equally strong sense that the influence of a new tech can be strong, subtle and unexpected. In other words, even if determinism in its strongest form isn’t correct, there’s a reason we keep arguing about it… in many cases, tech’s the trigger for a new cycle of behavior.

I’ve been arguing for some time now that language is a critical force that shapes how we do and don’t use the internet. I made the case a year ago that we need sustained, distributed human efforts to make translation more pervasive in a polyglot internet. In recent talks, I’ve been offering a thought experiment – how would our online behavior change if we never encountered a webpage in a language we don’t speak? What if our browsers were smart enough to know what languages we were comfortable reading and when they encountered another tongue, they’d either find a human translation or use an automatic translator to provide a comprehensible (if imperfect) version of the page?

My friend Brian McConnell of WorldWideLexicon – now Der Mundo – has taken a major step towards making this a reality. Brian is one of the pioneers helping think through distributed translation on the internet. Der Mundo is a long-standing project that invites anyone to help translate interesting texts, contributing whole or partial human translations of documents, allowing users to discover translations of documents and server administrators to make their sites open for translation. Brian has now taken a critical new step of introducing a Firefox plugin which determines the language of a page you’re browsing, looks for human translations and offers them to you, and otherwise connects with tools like Google Translate to provide you with comprehensible text.

The plug-in is part of a family of tools that Brian identifies as necessary to bridge online linguistic barriers: browser tools, web server tools (to open sites for translation), global translation memories (letting translators cooperate and share accumulated knowledge) and language service providers (services that can provide rapid, if imperfect, translations of texts between language pairs). The translation plugin is in an early stage – and can be a bit unstable at present – but it’s thrilling to see someone who’s thought so thoroughly about this field take the steps to bring theory into practice. In his essay on the tools, Brian offers his hope that we’ll see translation integrated into browsers like Firefox in the future. From his words to Mozilla’s ears.

In the meantime, Brian is trying to raise money to accelerate and extend his efforts. If you’re interested in the challenges of multilingualism on the net and can lend a hand, either with fiscal contributions or by becoming part of a development team, Brian’s done very important work for a long time and is one of the smartest and most respected people in the field.

Translation integrated into the browser doesn’t guarantee that internet users will become more cosmopolitan and less parochial in their internet usage. But the absence of tools that make it easy to stumble upon, understand and become interested in content in other languages makes it much less likely that we’ll use the internet as a way to bridge cultural and lingustic barriers. I think of Brian’s project as one of many necessary, though not sufficient, conditions to let us achieve a future where we’re all exposed to great ideas online, no matter what original language they were written in.