Adults are faster language learners than children, at the start. Studies that give children and adults the same new material in a lab, or follow them through their first months of living in a new language, find that the adults do better, perhaps because they bring conscious strategies, and what they know about their first language, to the task. The children’s advantage shows up only over the years that follow.¹

So how to learn a language as an adult is less a race against a biological clock than a question of hours and how you spend them. Adults can reach a high, professional level in a new language, but native-like grammar is rare for people who start after childhood, and the best-supported route mixes understanding, speaking and deliberate study rather than relying on any single method. Five steps, most of them backed by meta-analysesmeta-analysis: A study that combines the results of earlier studies on the same question into one overall estimate. Pooling makes the estimate more precise, but it cannot repair the studies it pools: a meta-analysis of surveys is still survey evidence.Full entry in the glossary:

1. Take in plenty of language you mostly understand
2. Speak and write early, and ask to be corrected
3. Study some grammar on purpose
4. Review new words by testing yourself, spaced out
5. Treat an app as one tool, not the whole plan

Adults learn fast at first, but native-like grammar is rare

Adults can learn a new language well, and start faster than children, but people who begin after childhood rarely end up with grammar indistinguishable from a native speaker’s. A very large 2018 online test of English grammar by Joshua Hartshorne, Joshua Tenenbaum and Steven Pinker suggests why: the ability to learn grammar declines in the late teens, and late starters run out of high-speed learning time.¹

Myth

After childhood, the window for learning a new language has closed.

Fact

Adults learn quickly, especially at first. What becomes rare after a late start is grammar and an accent that pass for a native speaker's.

The study separates two things most arguments about age run together: how fast you learn, and how far you eventually get. The authors use a race as the picture. That two runners both finished tells you each started early enough and ran fast enough, not that they ran at the same speed.¹

The study

Moderate evidence

A viral grammar quiz used to chart learning ability across the lifespan

The researchers built a short grammar quiz that spread widely online, then used a model to separate a person’s current age, the age they first met English and their years of experience. Grammar-learning ability stayed high until about 17.4 years of age and then declined steadily. To reach native-level grammar, learners needed to start by about 10 to 12, because later starters had too few years at the high rate before it fell.¹

Why the years matter: in the same data, even native speakers kept improving until about age 30, so mastering grammar takes many years even at the fastest rate. By the study’s estimate, a learner who starts at 30 has fewer of those years and a slower rate. The finding covers written English grammar only; the authors note that vocabulary and pronunciation may follow different timetables.¹

The shape of the decline is argued over. A 2022 reanalysis of the same data by Frans van der Slik and colleagues argued that for several groups of learners it fits a steady decline better than a sharp drop, and that the age effects may reflect schooling and living circumstances rather than a built-in developmental limit.² Hartshorne replied in 2024 that the reanalysis did not actually test sharp against gradual, and that corrected analyses confirmed a sharp drop in late adolescence.³ Neither side claims that adults cannot learn; the dispute is over why late starters end lower on average, and how suddenly.

Think of a 40-year-old engineer moving to Lisbon. In the first months she will likely out-learn her 8-year-old, reading signs and following meetings sooner. Ten years on, the child will probably sound local and she probably will not, though she can still run those meetings in Portuguese.

The practical move is to write your goal as something you can do (hold a meeting, read a contract, chat with in-laws) rather than as how native you sound. A goal like that also tells you what to practice first. For the general science of study behind the steps below, see the learning how to learn guides.

Further reading

• 

  How Languages Are Learned

  by Patsy M. Lightbown and Nina Spada

  Two researchers explain the age question, input, feedback and grammar teaching in plain terms, the same evidence this article weighs.

As an Amazon Associate WiserHours earns from qualifying purchases.

How to learn a language as an adult: five steps

The steps come from research on second-language learning, most of it with classroom students. The first two supply raw material, the next two make it stick, and the last puts apps in their place.

1. Take in plenty of language you mostly understand

You learn a language mainly from messages you can follow. Stephen Krashen’s influential input hypothesis, set out in 1982, holds that we acquire a language by understanding input that is a little beyond our current level, with context filling the gaps.⁴ His stronger claims, that input alone is enough and speech emerges by itself, are contested by the research in the next two steps.

“Mostly understand” means most of the words. An influential 2000 study put the figure at about 98 percent of the words in a text for adequate comprehension, but a 2023 replication with 104 adult learners in Sri Lanka could not fully reproduce that result.⁵ The exact threshold is uncertain; the direction is not. If you reach for a dictionary every sentence, you are decoding rather than reading.

Extensive reading, meaning a lot of easy, self-chosen reading, is one form of input with enough studies behind it to pool. A 2025 meta-analysis by Nina Sangers and colleagues found small to medium positive effects on reading, vocabulary, writing, speaking and general proficiency, with larger effects when learners chose from a limited set of texts and had some form of accountability.⁶ In an earlier meta-analysis, adult readers (college students and older) showed larger reading gains than children and teenagers, a comparison across studies rather than a test within one.⁷

In practice, that might be a graded reader for learners on the train, and a two-line summary sent to a friend or tutor at the end of each chapter. The summary is the accountability, and the short list of graded titles is the limited choice.

2. Speak and write early, and ask to be corrected

Reading and listening build understanding, but producing the language seems to need separate practice. A 2013 meta-analysis by Natsuko Shintani, Shaofeng Li and Rod Ellis compared grammar teaching built on comprehension with teaching built on production: learners made large gains with both, but for the ability to produce the forms, production-based teaching did better on delayed tests.⁸

It is the familiar gap between following a podcast in your new language and freezing when you try to order lunch. Speaking early is practice at the second.

Mistakes are useful if someone points them out. A 2010 meta-analysis of 33 studies by Shaofeng Li found a medium overall effect of corrective feedback that held over time, with larger effects in lab studies than in classrooms.⁹ A friend who politely understands your mistakes gives you conversation; a tutor who corrects them gives you something to fix.

3. Study some grammar on purpose

Studies that point learners to grammar and form report gains. A 2019 meta-analysis of 54 studies by Eun Young Kang, Sarah Sok and ZhaoHong Han found a large overall effect of instruction.¹⁰

How explicit that attention should be is less settled. A 2000 meta-analysis by John Norris and Lourdes Ortega, pooling experimental and quasi-experimental studies, found explicit instruction more effective than implicit instruction.¹¹ The 2019 review, with newer studies, found only a minor difference between the two.¹⁰ In both reviews, results depended partly on how the outcomes were measured, and neither shows how grammar study adds up over years of real use.

A reasonable reading for a self-taught adult: when a pattern keeps tripping you up, look up the rule, try a few sentences with it, then go back to reading and conversation where you will meet it again.

4. Review new words by testing yourself, spaced out

Vocabulary is where general learning science applies most directly. The two habits with the strongest evidence, recalling from memory and spreading reviews over days and weeks, are covered in why self-testing and spaced review beat rereading. Recalling a word feels harder than looking at it again; that guide explains why the effort pays off.

Say a chapter gives you a word you keep meeting. Put it on a card with the sentence you found it in, and try to recall the meaning before you flip it: the next day, again a few days later, then after a week or two. Cards you miss come back sooner. You do not need to match any of this to your supposed learning style either.

5. Treat an app as one tool, not the whole plan

App-based study looks helpful, but the studies behind it are weak. A 2022 meta-analysis of 23 studies by Mariela Mihaylova and colleagues found that learners using language apps scored higher than learners taught with traditional pen-and-paper methods, by a moderate-to-strong margin. The authors also found a high risk of biasrisk of bias: A rating of how much room a study's design, conduct and reporting leave for its result to be systematically wrong. It judges the method, not the motives of the researchers, and a high rating means the result may be off, not that it is.Full entry in the glossary and low-quality evidence across every study; only a handful said participants were randomly assigned. Most of the learners were university students, and nearly all were learning English.¹²

An app’s strength is contact: ten minutes of word practice on a commute is easy to keep up every day. What it rarely gives you is long stretches of reading, open conversation and a person correcting you. So let it carry the daily word work, and keep steps 1 and 2 outside it.

WiserHours

A week of adult language learning

• A goal stated as something I can do in the language
• Some reading or listening every day that I mostly understand
• At least one conversation or piece of writing with correction
• One grammar point looked up and practiced when it kept tripping me up
• New words reviewed by recall, on a widening schedule
• App time counted as practice, not as the whole plan

Find this checklist online:
https://wiserhours.com/learning/learn-a-language-as-an-adult/#checklist-learn-a-language-as-an-adult-week

Plan in hundreds of hours, not weeks

The most concrete time estimates for adults come from the US State Department’s language school, the National Foreign Affairs Training Center (formerly the Foreign Service Institute), which trains US government staff. It expects its full-time students to need 24 to 30 weeks for languages close to English, such as Spanish and French, and 88 weeks for Arabic, Chinese, Japanese and Korean.¹³ Between those, it expects about 36 weeks for German, Indonesian and Swahili, and about 44 weeks for a long list that includes Russian, Polish, Turkish and Vietnamese.¹³

Those timelines are for a specific level: a combined score of 3 in speaking and listening on the US government’s Interagency Language Roundtable scale.¹³ The scale calls level 3 general professional proficiency: enough to take part in most formal and informal conversations on practical, social and professional topics, with noticeable imperfections that virtually never get in the way of understanding.¹⁴

1. Category I, 24 to 30 weeks: languages close to English, such as Spanish, French, Italian and Dutch
2. Category II, about 36 weeks: German, Haitian Creole, Indonesian, Malay and Swahili
3. Category III, about 44 weeks: a long list including Russian, Polish, Greek, Hindi, Turkish and Vietnamese
4. Category IV, 88 weeks: Arabic, Cantonese, Mandarin, Japanese and Korean

Full-time means about 23 hours of class and 17 of self-study a week, to reach level 3 in speaking and listening. Source: US State Department language school (formerly the Foreign Service Institute).

The weeks are intensive. The school says a typical week is 23 hours in class and 17 hours of self-study, and that actual time varies with a learner’s natural ability, earlier language learning and time in the classroom.¹³ Its students also study full-time with native or near-native instructors, conditions few self-taught adults can match.

To run the same sum for another group of languages or a different weekly pace, change the inputs below; the planner starts at an hour a day for the group that includes Spanish.

For planning, judge progress over months, not weeks, and choose the language with its distance from English in mind. Japanese asks for nearly three times the weeks of Spanish, so by the same ratio a target that fits Spanish in two years could take about six for Japanese.

What the evidence supports, at a glance

The best-supported approach is a mix of understandable input, production with feedback and focused grammar study, each backed by meta-analyses of classroom and lab research; app studies are promising but low in quality.

The common limit: most of this research tests students, many learning English, on targeted measures within weeks. How these methods add up over the years an adult actually spends on a language is largely unmeasured. That is a reason to distrust any single method sold as the answer, and to measure yourself the way the government’s scale does: by what you can do in the language.

The bottom line

Starting a language as an adult costs you the easy route to a native accent and native-like grammar, not the ability to become fluent. Plan in hundreds of hours, spend them on reading and listening you mostly understand and on conversation where someone corrects you, look up grammar when it gets in your way, and let an app handle the daily word practice rather than the whole job.