Partage :

Translation as a method, crowdsourcing as a tool

Modern digital platforms welcome translation as an effective language-study technique that combines learning with collective contributions to business and community projects around the world.

By Gulnara Shaydullina

Once a solitary intellectual exercise, translation in the Web 2.0 era is often a collaborative effort of professionals or a collective pastime for amateurs. Technology facilitates human intellectual collaboration and creates new business models, crowdsourcing being one of them. A product of the digital revolution, the crowdsourcing of translations gives rise to new phenomena in all spheres of life, including online language learning.

In respect to language acquisition, translation is argued to be the fifth macro-skill—in addition to speaking, listening, reading, and writing—and it was widely used as a learning method. Several decades ago, though, it was replaced by the communicative approach, still a leading trend in formal foreign language teaching.

The online language-learning classroom, however, reclaims translation as an effective and motivating teaching method, offering students not only challenging exercises, but also a sense of contributing to the common good. A number of language learning websites—Languages on the Web and Translations for Progress among them—offer volunteer translation opportunities as a way to perfect language skills and to contribute to a good cause at the same time. New projects are being developed in various languages, although the concept of learning a language through crowdsourced translation is only beginning to take shape.

Teaching languages through crowdsourcing

Combining translation as a teaching method with participation in crowdsourced projects seems like a winning combination for both businesses and non-profit organizations, but software developers and lesson-plan designers have yet to discover a formula for a sustainable process. In this respect, the successes and mistakes of one of the trend’s leaders can teach those who follow. Duolingo, while not being the first educational online enterprise that tried using crowdsourced translation as a means of teaching, it is certainly one of the most studied ones. In 2013, committed to providing free language tuition, it managed to bring in partners, BuzzFeed and CNN, to finance its operations in exchange for translated content. The Duolingo creators aspired to organize its millions of users speaking Spanish, Brazilian Portuguese, and French to translate online news into their mother tongue while studying English, thus turning the learning process into a massive-scale online collaboration tool.

Combining language learning with crowdsourced micro-tasking—translation of text fragments and sentences—as an exercise, Duolingo set an example of how digital technologies can bring together traditional forms of teaching with new ways of doing business: it offered language learning with translation as its by-product.

What sets Duolingo’s approach apart is that translation was offered even for beginners, who started by translating very simple sentences into their mother tongue. There was a game element to it, too: the program was designed to provide instant feedback and let learners gain “skill points” all the while encouraging peer-to-peer collaboration and competition. The elements of challenge, the points systems, and the reputation gained among peers were there to boost motivation. The translation process also involved peer editing, an example of human computation filling the gaps where computers still fail. Quality assurance was therefore based on crowdvoting, the principle used in other crowdsourced translation projects by major online players—Facebook and Twitter.

Toward a more effective business model

Eventually, despite the good buzz and enthusiasm of users, Duolingo discontinued the crowdsourced news translation element of its language learning process. According to the platform creators, the business model did not prove to be competitive. The project’s critics also mentioned that controlling all the variables in texts for advanced learners was problematic, which led to a higher risk of wrong feedback; the way vocabulary and grammar were presented disregarded what corpus linguistics has to offer; and the quality assurance issues needed to be resolved.

Duolingo’s model might not have been a winning one for a particular business, but the human collective mind is still looking for a way to direct efforts toward the common good. That is why new online learning platforms are often designed as crowdsourced and bidirectional: while thousands of students learn languages through translation, numerous organizations benefit from their material being available in more languages, the audience in various corners of the world gains access to the information and, as a consequence, creates new content, which in turn, also needs to be translated.

Imperfect as they are, such projects are being developed, and their full potential is yet to be discovered. Donc, à suivre…

Gulnara Shaydullina

Gulnara Shaydullina is a Ph.D. student in Cognitive Informatics (Université du Québec à Montréal). Born and raised in Russia, she graduated from the Romance-Germanic Philology faculty at Bashkir State University (BGU, in Ufa, Russia) and worked as an ESL teacher, translator, and interpreter. Following her arrival in Canada, she received a Graduate Diploma in Translation (McGill University) followed by a Masters in Translation (Université de Montréal). Her current research interests are crowdsourcing in translation, technology and translation, e-learning, and teaching minority languages.


Partage :