IF YOU’VE been on Facebook recently, and you’re the kind of person who reads this column, there’s a good chance you’ve seen one of two viral language tests going round. Like many things on Facebook, they’re a fun diversion. Unlike many things on Facebook, they’re helping serious researchers learn about language and the mind.

The first is called “Which English?”, from the website Games With Words. Like another recent popular dialect quiz, it attempts to guess where you might have learned most of your English. But the version going round last year focused only on American dialects, and mainly asked questions about vocabulary and pronunciation. Which English? asks grammar-focused questions instead. It is also international: it will not only rank three guesses for the English dialect that has influenced you the most. It will also give three guesses as to your native language. Joshua Hartshorne, the MIT researcher behind the Games With Words lab that created the quiz, says that the top three guesses included the correct one about 80-90% of the time. Anecdotally, Johnson’s Facebook friends find it either eerily accurate or amusingly inaccurate. (It nailed your columnist’s own home dialect.)

What’s interesting is what’s going on behind the scenes. Mr Hartshorne, a cognitive scientist, conceived Which English? to research not dialect but language acquisition, and in particular the “critical period” hypothesis. Many linguists believe that there is a period for language learning (just before puberty) after which it becomes much harder to learn a foreign language.

Mr Hartshorne was not sure. The critical period may apply to the acquisition of a native-like accent, he thinks. But it may not apply to grammar. A previous study has suggested that grammar-learning ability falls off smoothly (not suddenly, after the critical period) depending on the age at which the learner begins learning. The earlier study relied on self-reporting, so he set out to find out what he could by actually testing people’s grammar, and then asking a few basic demographic questions. With around 500,000 test-takers so far, his results should be statistically robust, though he has gotten fewer non-Anglophone immigrants to Anglophone countries than he would like. (So do take the quiz, especially if you belong in that category.)

The second “language test” going around social networks uses language as a predictor of personality. Last year, researchers published a study in PLoS One, an online journal, in which 75,000 volunteers allowed their Facebook postings to be analysed, and also took a personality test on the “big five” personality traits: agreeableness, conscientiousness, openness, extraversion and neuroticism. Rather than hypothesise words or phrases that would predict personality traits, the researchers let the data speak for themselves, seeing which words emerged most typically with which traits. The word-clouds emerging from the study, for low and high values of each of the five criteria, are fascinating. Disagreeable people swear, introverts like anime, conscientious people invoke prayer and blessings (their word-cloud is pictured above), and those less “open” (ie, those who are unalytical and unreflective) are more likely to use internet-shortenings (UR, 2day, any_1) and omit punctuation (dont, wont). More.

See: Johnson | The Economist

Subscribe to the translation news daily digest here. See more translation news.