Every language app will tell you how long it takes, and the number is always a little too good. Ours has to produce one as well — at the end of onboarding it names a month — so it is worth writing down the arithmetic behind it, including the parts that make the answer worse than you want.
The number everyone quotes
The US Foreign Service Institute publishes rough training times for English speakers learning other languages. The figures move around between editions, but the bands are stable:
| Category | Languages | Classroom hours |
|---|---|---|
| I | Spanish, French, Italian, Portuguese, Dutch, Norwegian | roughly 600–750 |
| II | German, Indonesian, Swahili | roughly 900 |
| III | Russian, Hindi, Greek, Turkish, Vietnamese, Thai | roughly 1,100 |
| IV | Japanese, Korean, Mandarin, Cantonese, Arabic | roughly 2,200 |
Those numbers get repeated constantly and almost always misused. Four things about them.
They describe full-time intensive classroom instruction, typically twenty-five hours a week of class plus several hours of homework a day, with professional instructors and small groups. Nothing you are doing resembles this.
They describe selected adults, screened for aptitude and highly motivated, being paid to be there.
They target professional working proficiency — able to do a diplomat's job in the language. That is an extremely high bar, far above what most people mean by "learn a language".
And they are hours of instruction, not hours of exposure. The homework, the reading and the living-in-the-country part sit outside the number.
What the FSI table is genuinely good for is the ratio. Japanese takes an English speaker roughly three times as long as Spanish. That part holds however you study, and it is the single most useful thing in the table.
Why your number is different, in both directions
Worse: you are not doing twenty-five hours a week, you do not have an instructor, and your attention is divided.
Better: you probably do not want professional working proficiency. Following a conversation, watching video without subtitles, reading comfortably, and holding your own on holiday is a much lower target, and it arrives far earlier on the curve.
Worse again: spaced-out study is less efficient per hour than intensive study. Six hundred hours over two years is not six hundred hours over six months. You lose ground between sessions and spend part of each one recovering it.
Better again: there is more to work with now. Access to unlimited native material, transcripts on demand, and the ability to look a word up in a second are real multipliers that the 1970s language lab did not have.
The honest summary is that nobody can give you a number that is true for you. What anybody can give you is a consistent piece of arithmetic that you can sanity-check and adjust.
Ours, written out
We estimate progress in hours of listening, not days of streak, because listening hours are the thing that actually moves. The rough figures we use for how long a rung takes, at the pace of somebody doing it through real content:
| From | Hours of listening to the next rung |
|---|---|
| Starting from zero | about 60 |
| Some basics | about 90 |
| Conversational | about 150 |
| Advanced | about 200 |
Then two corrections. Not every minute of a daily goal is listening — you pause, you read the line again, you look at a word — so we count 80% of the goal as real listening time. And we round up rather than down at every step, because a target you beat is a reason to stay and one you miss is a reason to leave.
The arithmetic is then trivial, and you can do it for yourself:
days = (hours_for_your_rung × 60) ÷ (daily_minutes × 0.8)Fifteen minutes a day, starting from zero: 60 × 60 ÷ 12 = 300 days. Ten months to get from nothing to comfortable with common words and phrases in a Category I language.
Thirty minutes a day: 150 days. Five months.
Ten minutes a day: 450 days, and that is the number worth sitting with, because ten minutes a day is what most people actually do.
What the arithmetic will not tell you
Those hours assume comprehensible hours. An hour of video you understand nothing of does not count. This is the one input to the calculation people get wrong by the largest margin — the hours are hours of understanding, and material far above your level converts at a fraction of face value. That is the whole argument for grading content before serving it, and it is why finding material at your level is not a nicety.
Hard languages multiply the hours, not the days. If Japanese is three times Spanish by the FSI ratio, the rung figures above stretch accordingly. The writing system is a large separate cost that sits on top and does not overlap with listening at all.
Progress is not linear and does not feel like the graph. You will have a month where nothing improves, then a week where something obviously has. The plateaus are real and they are where most people quit, usually about four months in.
Consistency beats intensity by a wide margin. Twenty minutes every day for a year beats three hours every Saturday, and it is not close. The mechanism is not motivational — it is that retrieval strength decays, and the interval between exposures is what decides how much of each session is recovery rather than progress.
The version of the question worth asking
"How long to learn Spanish" has no answer. These do:
- How long until I can follow a video made for native speakers, with a transcript? Sooner than you think. Weeks, in a Category I language, if the clips are chosen to be at your level. This is the milestone worth chasing first because it is the one that makes everything after it sustainable.
- How long until I can follow it without the transcript? Months.
- How long until I can follow two native speakers talking to each other, at speed, with slang? Years, for any language, and this is the step almost everybody underestimates. See why native speech is a wall.
- How long until I can read a novel? Somewhere in between, and much sooner than the listening equivalent, because you control the pace.
The first of those is the one we build the plan around, and the reason is not marketing. Sixty hours at ten minutes a day is fifteen months, and asking somebody to commit to fifteen months before they have watched one clip is asking for a decision nobody can honestly make. The first real milestone is weeks away, it is the one they will feel, and it is the one that decides whether the fifteen months ever happen.
One number to keep
If you want a single rule of thumb, this one survives contact with reality better than most:
An adult who understands twenty minutes of real material a day, every day, in a language close to their own, will follow ordinary conversation within a year and be genuinely comfortable within two. Triple both figures for Japanese, Korean, Mandarin or Arabic.
Everything else is detail — and the detail that matters most is whether the twenty minutes are comprehensible, not whether they happened.