After decades of funding of European research in language and translation technology, the new European Commission wants to turn off the money tap. European researchers are staggering and they wonder why. One theory is that the politicians feel it is money wasted. Despite a diligent European investment program in machine translation, US corporations seem to have won that battle.
Let Google and Microsoft run with it, you hear Andrus Ansip thinking. Ansip is Vice President of the European Commission responsible for the Digital Single Market. He’s got a lot more on his plate. He will be held accountable for a rapid increase in cross-border online shopping: to 20% of the European population by the end of 2015 (compared to 12% by the end of 2013). How he’s going to do that? By investing in high-speed broadband internet and in digital service infrastructures. That will do the trick. The fact that the population of 300 million speak at least 24 different languages and are stubborn about that is not seen as a real issue. Or is it?
We don’t know. Despite the transparency displayed on the sites of the European Commission about budgets, meetings and agendas, it is hard to discover how decisions and policies are being reached. No doubt that hopes and fears and pride all come into play. In this case – another theory among the researchers who see their life’s work being threatened – fear and pride win over hope. Which politician, in his right mind, would want to hazard his career over a problem that cannot truly be solved within a few years, i.e. getting computers to produce high-quality translation?
It may be plain pragmatism (let Google and Microsoft run with it) or it may be fear (can’t be solved), but what is not acceptable is ignorance. Everybody knows that foreign languages remain a barrier to communications (can’t read, won’t buy) even if we have fast and affordable internet. And, what’s more, we cannot ignore the founding principle of the European Union that we are a multilingual society. A promise was made to citizens in all member states that they will always be able to speak their own language and be part of the European Union.
So what happened? Where did they lose it? Diligent and dedicated funding in ambitious language and translation research projects like Eurotra, Verbmobil and Moses over the last two decades created tremendous brain waves in Europe. When it was time to bring the ideas to market and apply all this intelligence, the takers were hard to find. Apart from some small translation technology start-ups that were spun off from the research projects, not very much came out of it. Big companies did not dare to burn their fingers on it
The Europeans have the brains, but lack the guts. In American business culture, the language barrier is simply another IT problem. You throw some code, data and dollars at it, and you’ll be able to solve it. Or maybe not. The only way to find out is to try and then ask users what they think of your Minimum Viable Product. Failure is the road to success. For Europeans language is more than an IT problem. They take it personally. Many of them still have to overcome the shock that hundreds of millions of users accept imperfect machine translations. Europeans have a long history with their own languages. Deep in their heart they may feel resistance to letting it be deciphered and unlocked by computers.
This is exactly why – that special relationship that the Europeans have with multilingualism – we think that the European Commission cannot give up researching into the possibilities of a European population freely communicating in 24 different languages. Sixty years of research in machine translation has brought us this close. Imagine how successful the Digital Single Market would be and how the economy would grow if every European citizen could simply speak his own language and understand all his fellow citizens. European citizens would benefit everywhere.
In our MT Market Report (published in August 2014) we describe how translation is becoming a utility embedded in every screen, every app, every device and now every 'thing.' Machine translation technology is a force multiplier and a catalyst for innovation in many different aspects and sectors of life and business. In a follow-up Moses MT Market report – just published – we describe how Moses open-source MT technology has driven competitiveness and innovation in the sector. Moses was funded for over ten years by the European Commission and now constitutes 20% of the global machine translation sector.
With or without funding from the European Commission the goals will be achieved. Andrew Joscelyne reports in the latest issue of TAUS Review that eBay plans to sell all of its inventory all across Europe by deploying machine translation. They expect their cross-border trade to grow by 30% in the next few years. Besides Google, Microsoft, eBay and Facebook we see many more small and large companies standing up ready to help us communicate easily across languages. With or without EC funding the world will overcome the language barriers. It is plain wrong that Europe is withdrawing now, at this ‘moment suprême’.
Now, surely, is the time for a big idea – for Europe and America to work together to make it happen. Time to launch the Human Language Project. The Human Language Project – analogous to the Human Genome Project – will collect all the data for most languages spoken in the world. Just as the shared, open repository of human DNA data spurred unprecedented growth in the life sciences industry and gave rise to great medical discoveries, the Human Language Project will be the foundation for major steps forward in education, well-being and prosperity for all. (See the article and the many comments related to the Human Language Project: “It’s time for a Big Idea”)
What it takes is for the brains of Europe and the guts of America to unite and work together. An investment of 1 billion Euros (less than 1% of the budget of the European Commission) matched by 1 billion from the Americans, would be enough to get the project going.
We invite business leaders and politicians to discuss the following sorts of questions:
- Whether we should let Google and Microsoft run with it, or invest in our own translation technology;
- What governments can do for the translation industry;
- Translation leaks. The upside and downside of sharing data;
- Treating translation quality metrics as business intelligence;
- Bridging the gap between business and education.
Share your comments below. After all, that is what this blog is for, to stimulate a most important discussion.