DuHocNgonNgu.comDu Học Ngôn Ngữ
The science of language acquisition

The sensitive period for learning a foreign language: is 8–17 too late, and what does the research say?

Lenneberg's sensitive period hypothesis, the Johnson–Newport and Hartshorne 2018 studies, and what they really say about children aged 8–17 learning English or Chinese.

The sensitive period for learning a foreign language: is 8–17 too late, and what does the research say?

A child aged 8–17 is not too late to become good at a foreign language. What research on the sensitive period shows is this: the later a learner starts, the harder it is to reach an accent and a feel for grammar identical to a native speaker's, but the ability to reach fluent, working proficiency for study, work and friendship remains fully intact. On many measures, teenagers actually learn faster than young children in the first few months.

The debate lies elsewhere: is there a "window" that slams shut at a particular age, or just a gentle slope; and is the cause the biology of the brain, or the amount and quality of language exposure, which change with age? This article walks through the main studies in chronological order, explains what each one measured, and draws out what parents can do with a child of this age.

How the hypothesis began: from Penfield to Lenneberg

The idea that a child's brain learns language more easily than an adult's already appears in Speech and Brain Mechanisms, published in 1959 by the neurosurgeons Wilder Penfield and Lamar Roberts. They observed that children with damage to the language areas of the left hemisphere usually recovered speech better than adults with similar damage, and concluded that the child's brain was still "plastic".

In 1967, the linguist and neurologist Eric Lenneberg turned this idea into a fully formed hypothesis in Biological Foundations of Language. In his view, the ability to acquire language naturally lasts from about the age of two until puberty, and is tied to the specialisation of the two brain hemispheres. After puberty, language can still be learned, but it takes more conscious effort and is rarely complete.

Two things are worth remembering when reading Lenneberg. First, he was mainly discussing the mother tongue, the first language, not a foreign language learned later. Second, his argument about hemispheric specialisation was challenged early on: studies in the early 1970s showed that specialisation happens much earlier than puberty, so the mechanism he proposed is no longer accepted as he stated it. What remains is the question he raised: how does the age at which someone starts affect the outcome of language learning?

Johnson and Newport 1989: the study that carried the hypothesis over to foreign languages

The most-cited study for foreign languages is the paper by Jacqueline Johnson and Elissa Newport, published in Cognitive Psychology in 1989. Its design and main findings were as follows:

  • Who was tested: Korean and Chinese speakers who had moved to the US at different ages, from about 3 to nearly 40, and had lived there for many years.
  • What was measured: grammaticality judgement, that is, hearing an English sentence and saying whether it is grammatically correct or not, across many structures such as tense, plurals, articles and word order.
  • Findings: those who arrived in the US before about age 7 performed almost like native speakers; from there, scores fell steadily with age up to puberty; those who arrived after puberty scored lower and varied enormously from one person to another.
  • What is often overlooked: among late arrivals, age of arrival no longer predicted the outcome well; individual differences were far larger. In other words, after puberty other factors such as motivation, amount of exposure and learning approach start to matter more.

In 2001, David Birdsong and Michelle Molis repeated this design with Spanish speakers and found no clear break at puberty: scores kept declining gradually into adulthood, and some late arrivals still reached very high levels. Part of the reason may be that Spanish is closer to English than Korean and Chinese are, which is worth noting for Vietnamese speakers, whose language is distant from English.

The 2018 study of more than 600,000 people: age 17 and age 10

In 2018, Joshua Hartshorne, Joshua Tenenbaum and Steven Pinker published an analysis in the journal Cognition based on an online English grammar quiz taken by around two-thirds of a million people, together with information on when they started learning, how many years they had learned and where they lived. This scale was hundreds of times larger than any previous study.

The authors' conclusion has two parts. First, the ability to learn grammar appears to stay high until around the end of the teenage years, about age 17, and only then declines. Second, because it takes many years to reach native-like grammar, people who want to reach that level usually need to start before about age 10 in an immersion setting. These two points are often merged into "after 10 it's over", which is a misreading.

The study has also been criticised: the data are cross-sectional rather than following individuals over time; the quiz only measured grammar, not listening, speaking, reading or writing; and the results depend on the assumptions of the mathematical model. Even so, for parents of children aged 8–17 the main message is rather reassuring: at this age, your child is still in the period when grammar is learned most effectively, and what makes the difference is the amount of real language they are exposed to.

Learning faster and learning further: two different questions

A 1979 review by Stephen Krashen, Michael Long and Robin Scarcella separated two questions that parents often treat as one: who makes faster progress in the early stage, and who reaches a higher level after many years. The table below brings together what later studies found when comparing young children with teenagers who had the same exposure to the language.

QuestionYoung children (under about 8)Teenagers aged 11–17Reference
Progress in the first few monthsSlower in grammar and vocabularyFaster, thanks to reading ability, rule generalisation and working memorySnow and Hoefnagel-Höhle (1978), English speakers learning Dutch
Classroom learning of a few hours a weekVery small or no advantage from starting earlyCatch up and often overtake with the same number of hoursBarcelona Age Factor project, Carmen Muñoz (2006)
Accent after many years living abroadUsually close to nativeUsually a foreign accent remains, to varying degreesFlege, Munro and MacKay (1995), Italians in Canada
Grammar after many years of immersionOften reaches native levelReaches a high level, rarely identical to nativeJohnson and Newport (1989); Hartshorne et al. (2018)

The second row matters most for Vietnamese parents: the advantage of a young start only shows when there is a very large amount of exposure, the kind that comes from living in the language every day. Starting early with only a few hours a week, the advantage all but disappears, and learners who start later but study more intensively catch up quickly.

Why older learners find a native accent harder: competing explanations

The neurological explanation. The brain becomes less plastic with age, especially in the circuits that process sound. This view is supported by research on infants: from about six to twelve months of age, babies gradually become worse at telling apart pairs of sounds that do not exist in their mother tongue, as Janet Werker and Richard Tees described in 1984. But this concerns sound perception and does not explain all the differences in grammar.

The established mother tongue explanation. Adults hear a new language through the "filter" of the sounds and structures of their mother tongue. The longer someone uses Vietnamese, the more firmly the Vietnamese sound system is set, so English sounds close to Vietnamese sounds get "pulled" towards the Vietnamese ones. This is the central idea of James Flege's Speech Learning Model.

The environmental explanation. Young children who move abroad usually go to school and play with native-speaking friends all day; adults who move usually work and socialise within a community from their home country. The amount and quality of input differ hugely, and part of the "age effect" is actually the effect of the environment that comes with age.

The learning-style explanation. Young children learn implicitly, absorbing through exposure; teenagers and adults can also learn explicitly, analysing rules. Explicit learning helps early progress, but it does not automatically become a reflex. Researchers such as Robert DeKeyser argue that late learners rely more on analytical aptitude.

For children aged 8–17: what the research suggests parents do

Turning these findings into concrete decisions, parents can work through the following steps in order:

  1. Set the right goal. A reasonable goal at this age is to speak clearly, understand native speakers, and use the language for study and work, not to "make sure nobody can tell they are Vietnamese". A slightly different accent does not stop listeners from understanding your child.
  2. Count the real exposure. Count the number of hours each week your child actually listens to and understands the language and has to speak it, including outside class. This is the variable that research shows matters most at this age.
  3. Prioritise intensive periods. A few weeks of continuous exposure, where the language is a means of living rather than just a school subject, gives your child something a few classroom hours a week can hardly give: being forced to understand and forced to speak.
  4. Do not skip guided pronunciation work. Because accent is the hardest part for those who start after early childhood, a teacher who points out exactly which sounds your child is mishearing is worth more than many hours of passive listening.
  5. Keep it going after the intensive period. Long-term results depend on the years of accumulated exposure; one intensive summer followed by a complete stop will not keep much.

This order also explains why researchers today usually speak of a "sensitive period" rather than a "critical period": the ability to learn declines gradually rather than disappearing, and the environment largely decides how far that ability is put to use.

Sayings about the age of language learning worth reading again, correctly

"After puberty you can never become good" is false. Research shows it is hard to reach a level identical to a native speaker on every measure, not that it is hard to become good. In studies of late learners who reach near-native levels, such as those by Kenneth Hyltenstam and Niclas Abrahamsson's group in Sweden, some people pass as native speakers to native listeners, even though very detailed tests still find small differences.

"Children learn like sponges" is only true when the child is immersed in the language. In a classroom of a few hours a week, young children do not absorb it faster on their own; many forget almost everything once the exposure stops. "The window closes at age X" is also misleading, because researchers have not agreed on any such age, and the different parts of a language (sounds, grammar, vocabulary) seem to be affected by age in different ways.

Vocabulary is the area least affected by age: adults keep learning new words all their lives. As for Chinese, Vietnamese speakers have a particular advantage in their store of Sino-Vietnamese words, which helps them recognise the meaning of many academic words; this has nothing to do with age, and everything to do with the distance between the two languages.

Frequently asked questions

If my child only starts learning English seriously at 15, can they reach a level good enough for university in English?

Yes. The level needed to study at university in a language, roughly B2 to C1 on the CEFR, is one that a great many people who started as teenagers or adults have reached. What decides it is the hours of study and exposure accumulated over the next few years, not the starting age.

What is the difference between a sensitive period and a critical period?

A critical period implies a window with a clear closing point, after which learning hardly happens at all. A sensitive period says there is a stretch of time when learning is easiest, after which the ability declines gradually but remains. For foreign languages, most current evidence leans towards the second view.

Does the hypothesis apply to Chinese in the same way as to English?

The general principle is the same, but most of the research has been done with learners of English. In Chinese, tones are an area of sound that may be affected by age; Vietnamese speakers are already used to hearing tones, so they start from a different point than English speakers. There are not yet many large studies specifically on Vietnamese learners of Chinese.

If age does not decide much, why do people still advise starting children early?

Starting early gives more years of accumulation and a chance of reaching a near-native accent, provided the amount of exposure is large enough. The sensible advice is to start when good exposure is possible, not to start as early as possible at any cost with a few hours a week.

Why do some adults become very good at foreign languages, even with a near-native accent?

Studies of exceptional late learners usually find a few factors in common: strong analytical language aptitude and good memory for sounds, high motivation, many years of living and working in the language, and deliberate pronunciation practice. They are a minority, but they show that the biological limit is not absolute.

Is it better to send my child abroad to learn a language at 9, or to wait until 13?

Research on age cannot answer this for you. At 13, children are usually more independent, progress faster in the first few weeks and remember more; at 9, they pick up sounds more easily but need closer supervision. Which suits your child depends on how independent they are and on the plan for keeping the language going after the trip.

Need specific advice for your case?

We will contact you within 24 hours.

Request consultation now

Related articles

Need more help?

Leave your details and we will get back to you within 24 hours.