A1 to A2 is the stretch where a language stops being a set of memorized phrases and becomes a small system you can recombine. At A1 you can produce what you have rehearsed. At A2 you can build sentences you have never heard, about routine matters, and understand someone doing the same back to you. The 90-day plan below gets there by front-loading the highest-frequency vocabulary, fixing pronunciation early, and reaching readable text sooner than most beginner courses allow.
What A1 and A2 actually describe#
A1 is the level of fixed expressions and predictable situations: introducing yourself, asking where something is, filling in a form, understanding slow, clearly articulated speech when the speaker is helping you. The grammar underneath is thin, and that is by design.
A2 adds recombination. You handle short social exchanges, describe your background and immediate environment, and read short simple texts of the kind that appear on menus, timetables, and personal messages. The descriptors are still tied to routine matters, but you are no longer reciting. That difference is what the 90 days buy. For the full ladder and how the levels relate to each other, CEFR levels explained sets out all six.
Why frequency ordering matters most at the very start#
Vocabulary is not evenly distributed. A small set of words does an enormous amount of work in every text ever written, and the return on learning them collapses quickly as you move down the list. Paul Nation's research on vocabulary size and text coverage is the standard reference for this, and the approximate shape looks like the table below. The exact figures vary by corpus and by how word families are counted, so treat them as indicative.
| Word families known | Approximate coverage of general text | What that feels like when reading |
|---|---|---|
| 100 | ~50% | Every second word is a gap |
| 500 | ~65–70% | You can identify the topic and little else |
| 1,000 | ~80% | Roughly one unknown word in five |
| 2,000 | ~85–90% | Graded readers become genuinely comfortable |
| 3,000 | ~90%+ | Simplified news becomes readable |
The consequence for a beginner is direct: the first 1,000 words are worth more than the next 3,000 combined, so the order you learn them in is the single highest-leverage decision of your first three months. Learning because, already, and between before the names of zoo animals is not a preference. It is arithmetic.
This also means beginner courses that teach vocabulary by theme, in the order the syllabus finds convenient, are working against you at exactly the point where ordering matters most. Frequency lists are freely available for every language, and most well-built starter word packs are frequency-ordered whether or not they advertise it.
Days 1 to 14: sound and script before vocabulary#
Spend the first two weeks on how the language sounds and how it is written, with only incidental vocabulary. This feels like a delay and is not. Pronunciation errors set at the beginning are corrected slowly and imperfectly, and every word you learn with a wrong sound attached has to be relearned later.
The concrete work is small. Learn the writing system to the point of decoding, including any diacritics. Learn which letters and clusters do not say what they say in English. Learn where stress falls, and whether it is fixed. Then shadow short recordings, imitating rather than translating. For a non-Latin script, extend this block to three or four weeks and treat the script as the primary task.
Audio matters more here than later. Hearing a word pronounced by a native voice at the moment you meet it, rather than guessing from the spelling, is cheap to build on day one and expensive to retrofit on day two hundred.
Days 15 to 45: the first 500 words#
This is the bulk-acquisition phase. The goal is 500 high-frequency words in thirty days, which is roughly seventeen a day, and it is achievable because these words are short, common, and appear everywhere.
Two mechanics do the work. First, spaced repetition: introduce a word at short intervals and stretch them as you succeed, which counters the forgetting curve Hermann Ebbinghaus measured in 1885. Second, context. A word learned as an entry in a two-column list is fragile; the same word met inside a sentence carries its grammar and its typical company with it.
Attach an image and a personal mnemonic to any word that refuses to stick. Allan Paivio's dual coding theory (Paivio, 1971) is the reason this works, and dual coding for vocabulary covers how to do it without turning every card into a project.
Days 46 to 90: sentences, and the first material you can read#
Around 500 words, something changes: you can start reading rather than decoding, provided the text is chosen honestly. The first readable material is not a novel and not a news site. It is one of three things.
Graded readers written for A1 and A2, which control vocabulary deliberately. Short predictable texts where you already know the content, such as recipes, product descriptions, or a film synopsis you have seen. Or text in your own language with a share of its words swapped into the target language, so the surrounding sentence carries the meaning of each new word. That last approach is the diglot weave, first described by the linguist Robbins Burling in 1968, and it is what LingoBlend does: you paste something you already wanted to read, set how much of it appears in the target language, and tap any unfamiliar word for its meaning and grammatical form.
Whatever the source, the standard is the same. If you stop more than five times per hundred words, the text is too hard for building volume: drop a level or lower the density. Easy Spanish books for beginners collects starting points by difficulty, and the languages overview lists every pair available.
The 90-day schedule at a glance#
| Phase | Days | Daily time | Core activity | Cumulative words |
|---|---|---|---|---|
| Sound and script | 1–14 | 30–40 min | Pronunciation, alphabet, shadowing, ~50 incidental words | ~50 |
| Core vocabulary | 15–45 | 45–60 min | 15–20 frequency-ordered words daily plus spaced review | ~500 |
| Sentences | 46–70 | 45–60 min | Reading graded or blended text, 10–15 new words daily | ~800 |
| Consolidation | 71–90 | 60 min | Volume reading, review backlog, first spoken output | ~1,000 |
The review load grows even when the new-word count falls. By day 70 most of your session is maintenance, and that is the correct shape. A thousand words half-remembered is worth less than six hundred you can retrieve.
What 90 days actually produces#
Ninety days at 45 to 60 minutes a day is roughly 70 to 90 hours. Commonly cited guided-hour estimates put A2 at around 180 to 200 hours for a language close to your own. So: for Spanish, French, Italian, Dutch, or Swedish, 90 days lands you in late A1 or early A2. For German, add time. For Japanese, Chinese, Finnish, or Hungarian, expect roughly half as much measurable progress, with the script phase alone consuming the first month. How long it takes to learn German shows how the same hours land differently by language.
What the 90 days reliably delivers is a working core vocabulary, pronunciation that does not need unpicking, the ability to read graded material without constant lookups, and enough sentence patterns to survive a transaction. What it does not deliver is conversation. Speaking lags reading by months at this stage for almost everyone, and treating that lag as failure is the most common reason beginners quit in month four. The next stage, and the reading routine that carries it, is laid out in the A2 to B1 reading plan.