Window for learning second language may remain open longer than thought

It’s long been believed that young children become fluent in a second language more easily than their parents. A study now shows that the critical window for learning a new language is longer than previously believed.

Researchers from Boston College, Massachusetts Institute of Technology, and Harvard University found that the ability to learn the grammar of a new language fades around 17 years old, “at the crux of adulthood,” according to the study published in Cognition. Additionally, the study showed that achieving fluency like a native speaker in a foreign language requires starting to speak the language by 10 years old.

“The ability to learn the grammar of a new language fades around 17 years old.”

“This is surprising. These results have upended the discussions on this topic,” said lead author Joshua Hartshorne, Boston College assistant professor of psychology who conducted this research as a postdoc at MIT. “Previously we wondered, does the window for learning a second language start closing just after birth, at four or five years old, or at the beginning of puberty? There weren’t any hypotheses for the window closing at 17 or 18 years old.”

Data acquisition through machine learning and social media

The researchers posted a 10-minute online “Games with Words“ quiz called “Which English?” which tested non-native speakers’ grammatical proficiency in English.

The quiz began by asking participants their age and when they started learning English. Then, they were asked if sentences such as “Yesterday John wanted to won the race” were grammatically correct.

In addition, the quiz included questions to tease out which dialect the participant spoke. This allowed the quiz to become more of a game, as machine learning offered participants the “best guess as to which English you speak as well as whether your first (native) language is English or something else,” according to the site.

The quiz was widely shared on social media, attracting nearly 700,000 participants from around the world, making it the largest dataset for the study of language-learning ability.

Scientific and educational implications

Why the window closes at 17.4 years old remains unknown.“Could it be social, that you are in college or working and do not have time to study a new language? Could there be a neurological reason? ” Hartshorne asked. “That is for other researchers to investigate.”

However, the results raise several questions for educators.

Because people who started learning English before age 10 seem to learn much more than those starting later, classroom programs might be designed to start in primary grades. Typically, American school systems offer four years of a foreign language at the high school level.

“Educational systems have to consider what the goals for teaching language are. Is it to function in a language when traveling, or is it complete fluency?” he said. “If fluency is the goal, they might need to extend these offerings over a longer period of time.”

“Immersion experiences offer deeper language learning than classroom learning.”

However, Hartshorne emphasized that even those who start later than 10 years old can learn a great deal, and quickly. “The questions are: How good do you want to be, and how hard do you want to work at learning?” he added.

In addition, the study supports the theory that immersion experiences offer deeper language learning than classroom learning.

“The difference is enormous,” Hartshorne said. “People who had English classes but had never been to an English-speaking country maxed out at a much lower level than someone who has had a few years of immersion.”

The researchers have posted their data online for other scientists to study. “It is an enormous, rich data set and there is a lot we haven’t looked at,” he added.