Take a Quiz, Test a Language Algorithm
Like most people with an internet connection, you’ve probably taken an online quiz at some point to find out what “Star Trek” character you’re most like, or what city you should live in. Now, a cognitive scientist from MIT is capitalizing on the quiz craze to answer some serious questions about how we acquire language.
Joshua Hartshorne, a postdoc in MIT’s Computational Cognitive Science Group, developed “Which English?” as an effort to map English grammar around the world. Hartshorne and his team hope to learn three main things, according to the Games with Words blog:
1. How does the age at which you start learning a language affect how well you learn that language?
2. How is learning a foreign language affected by the language you already know?
3. How are the grammars of different English dialects related?
After completing the quiz, users are shown the algorithm’s best guess as to whether English is their native language, as well as what dialect of English they speak. Approximately 500,000 people have taken the quiz, and Hartshorne says the results are correct 80–90% of the time. He has created a visualization of the results so far.
Hartshorne believes that knowing someone’s language background has practical applications for education, localization of websites, and more. Follow his research progress.
Read more about Hartshorne’s project, as well as another language research project that’s crowdsourcing its data from Facebook, at The Economist.