How long does it take to learn a language: Foreign Service Institute (1947)
For an English speaker, the Foreign Service Institute (FSI) — the U.S. State Department’s diplomat-training school, founded in 1947 — estimates that reaching professional working proficiency takes roughly 600–750 hours of study for the easiest languages (Spanish, French, Italian, Portuguese, Dutch), about 900 hours for German, around 1,100 hours for “hard” languages such as Russian, Polish, Turkish or Hebrew, and about 2,200 hours for the four “super-hard” languages — Arabic, Chinese (Mandarin and Cantonese), Japanese and Korean. These figures come from more than seventy years of teaching adults in intensive, full-time classes, and they are the most widely cited answer to the question “how long does it take to learn a language?”
The number is a useful anchor, but it hides two things that change everything: what proficiency it measures, and how those hours are spent. The FSI target is a high bar — closer to CEFR B2–C1 than to holiday-phrasebook survival — and its hours assume a classroom pace almost no self-learner matches. Understanding both is the difference between a realistic plan and a discouraging myth.
Where these numbers come from
The FSI trains U.S. diplomats and foreign-affairs staff who need to work in another language, so it has a strong incentive to estimate training time accurately — a course that is too short sends unprepared officers abroad, and one that is too long wastes public money. Over decades of running these programmes, the School of Language Studies recorded how long its students actually took to reach a given standard, then grouped languages into difficulty categories based on that data.
Two features of the FSI setting explain why its numbers are both authoritative and easy to misread:
- The learners are motivated adults in intensive, full-time training. A typical FSI course runs about 25 hours a week of classroom instruction with small groups and native-speaker instructors, plus several hours of directed self-study every day. The published totals assume this pace — 600–750 hours is roughly six to seven months of full-time study, not a casual evening hobby.
- The goal is a specific, high level of proficiency. The hours are the time needed to reach “Professional Working Proficiency,” or level 3 on the Interagency Language Roundtable (ILR) scale in both speaking and reading (written “S-3/R-3”). ILR level 3 means you can work in the language — hold meetings, argue a position, understand the news — and maps to roughly CEFR B2–C1. It is far above the ability to order dinner or make small talk, which arrives much earlier.
So the FSI figure is not “how long to speak a language” in the loose sense most people mean. It is “how long, in full-time institutional training, to become able to do a professional job in the language.” Both parts of that sentence pull the number upward.
The FSI categories in hours
The modern FSI ranking uses four categories. Placement reflects how closely a language resembles English in vocabulary, grammar and writing system — the more shared ground, the fewer hours. Approximate figures:
| Category | Classroom hours (weeks) | Representative languages |
|---|---|---|
| I — closely related to English | 600–750 h (24–30 weeks) | Spanish, French, Italian, Portuguese, Romanian, Dutch, Norwegian, Swedish, Danish, Afrikaans |
| II | ~900 h (~36 weeks) | German, Indonesian, Malay, Swahili |
| III — hard languages | ~1,100 h (44 weeks) | Russian, Polish, Turkish, Hebrew, Hindi, Thai, Vietnamese, Greek, Finnish |
| IV — super-hard languages | ~2,200 h (88 weeks) | Arabic, Chinese (Mandarin & Cantonese), Japanese, Korean |
A few honest caveats about the table. The exact placement of some languages shifts between published versions of the ranking — Arabic in particular is sometimes shown in the “hard” group and sometimes in the “super-hard” one, and FSI marks a handful of Category III languages with an asterisk to note that they usually take longer than others in the same tier. German is often treated as a step above the other Germanic and Romance languages because of its more complex grammar. The categories are a summary of averages, not a fixed law: the point is the ratio — a super-hard language takes roughly three to four times as long as an easy one — more than any single hour count.
What this means for a self-learner
Almost nobody outside a diplomatic academy studies 25 hours a week. The useful exercise is to convert the FSI totals into the pace a real person keeps. Taking a Category I language at 700 hours and a Category IV language at 2,200 hours:
| Study pace | Category I (~700 h) | Category IV (~2,200 h) |
|---|---|---|
| 30 minutes a day | ~3.8 years | ~12 years |
| 1 hour a day | ~1.9 years | ~6 years |
| 3 hours a day | ~8 months | ~2 years |
| FSI pace (~25 h/week) | ~6–7 months | ~1.7 years |
Two things stand out. First, the totals are large because the target is high: reaching a comfortable travel-and-conversation level (roughly CEFR A2–B1) takes a fraction of the full figure — often a few hundred hours for a Category I language — so “I want to chat on holiday” and “I want to work in the language” are genuinely different projects with different timelines. Second, consistency beats intensity for most people’s lives: 30 minutes every day is 182 hours a year, and it is far more sustainable than a burst that collapses after a month. The FSI hours are best read as a distance, and your daily minutes set the speed.
What speeds learning up — and what does not
The FSI categories already encode the single biggest factor: how related the target language is to one you already know. A Dutch speaker learning German, or an English speaker learning Spanish, starts with thousands of near-transparent cognates and a familiar alphabet; a learner facing a new script and a tone system starts from zero on several fronts at once. Related research on cross-linguistic transfer confirms that shared vocabulary and grammar measurably lower the time to proficiency — which is exactly why the categories exist. Prior experience of learning any second language helps too, because the study skills transfer even when the words do not.
Beyond relatedness, the evidence points to a few things that genuinely help — and one popular assumption that does not:
- Distribution usually beats cramming for what sticks. The FSI’s intensive schedule reaches a test date quickly, but the wider research on the spacing effect shows that the same number of hours spread over time produces better long-term retention than the same hours massed together. Intensity front-loads exposure; spacing is what keeps it. For a self-learner not facing a fixed exam date, steady spaced review is the higher-yield strategy.
- Quality of practice matters more than raw exposure. Hours spent on active retrieval — trying to produce the language and getting feedback — build ability faster than passively re-reading or letting audio wash over you. This is the language-learning face of deliberate practice, and it is why a little difficulty is desirable: effortful recall strengthens memory in a way that easy recognition does not.
- Comprehensible, meaningful input raises the floor. Material you can mostly understand, embedded in real sentences and context, is retained far better than isolated word lists — a point that runs from Ebbinghaus’s meaningful-vs-nonsense finding to modern input research. Meaningful practice starts higher on the memory curve and decays more slowly.
- What does not reliably speed things up: chasing a shortcut that removes the hours. The FSI numbers are stubborn precisely because proficiency is cumulative — there is no method that turns 2,200 hours of Japanese into 200. Good methods make each hour count for more; they do not abolish the hours.
What this means for language learning
The FSI data is the most honest public answer to “how long will this take,” and its lesson is not discouraging so much as clarifying:
- Pick a target level before you count hours. Professional proficiency (ILR 3 / CEFR B2–C1) is a multi-hundred- to multi-thousand-hour project; conversational comfort is a much shorter one. Know which you want. The CEFR levels give you a concrete ladder to aim at.
- Turn the total into a daily habit. Because the numbers are cumulative, the only variable you control is hours per day. Thirty consistent minutes will get you there; sporadic marathons usually will not — and spacing those minutes out is what makes them last, the core idea behind spaced repetition in language learning.
- Make each hour count with meaningful, active practice. Learning whole sentences you can understand, and recalling them actively, extracts more from every hour than passive exposure — the principle behind learning with full sentences.
The FSI categories are estimates produced by one institution training one kind of learner toward one high standard — not a natural law that fixes your personal timeline. What they get right, and what a self-learner can borrow, is the shape of the problem: proficiency is a distance measured in hours, the distance depends heavily on how far your target language sits from the ones you know, and the reliable way to cover it is steady, meaningful, well-spaced practice.
Frequently asked questions
How long does it really take to learn a language?
For an English speaker, the FSI estimates roughly 600–750 hours for an easy language like Spanish or Dutch, about 900 for German, around 1,100 for a hard language like Russian or Polish, and about 2,200 for Arabic, Chinese, Japanese or Korean. Those hours target professional proficiency (about CEFR B2–C1). At 30 minutes a day, 700 hours is roughly four years; at an hour a day, about two. Basic conversational ability arrives much sooner, after a few hundred hours for the easier languages.
Why is Arabic or Japanese so much harder than Spanish for English speakers?
The FSI categories track how far a language sits from English. Spanish shares much of its vocabulary, grammar and alphabet with English, so a learner starts with a large head start. Arabic, Chinese, Japanese and Korean combine an unfamiliar writing system, sound system and grammar with almost no shared vocabulary, so several things must be learned from scratch at once — which is why they take roughly three to four times as many hours.
Can you learn a language faster than the FSI estimate?
You can reach the FSI hours in less calendar time by studying more intensively, and you can make each hour more productive with active recall, spaced review and meaningful input rather than passive exposure. What no method does is remove the hours: proficiency is cumulative, so the estimates are a floor for the amount of practice, not a schedule you can shortcut. Better methods raise the value of every hour; they do not delete them.
Sources
- U.S. Department of State, Foreign Service Institute — School of Language Studies. Foreign Language Training (category timelines and the ILR-3 proficiency target).
- Interagency Language Roundtable. ILR Scale — Speaking Proficiency (definition of Professional Working Proficiency, level 3); overview at Wikipedia: ILR scale.
- Foreign Service Institute. Foreign Service Institute — history and mission.
- Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354–380 (evidence for spacing over massed practice).
- Council of Europe. Common European Framework of Reference (CEFR) — global scale (proficiency levels used to interpret the ILR-3 target).