The words that matter most — why frequency decides what to learn first

Every language has hundreds of thousands of words. Nobody knows them all, including native speakers, and nobody needs to. What a learner needs is the right few thousand — and which ones those are is not a matter of taste. It has been counted.

The arithmetic of frequency

Take any large body of text — newspapers, conversations, subtitles — and count how often each word appears. The result is always the same shape: a handful of words used constantly, then a long tail used rarely.

In English, the 100 most frequent words make up roughly half of everything written. The most frequent 1,000 word families cover around 80 per cent of ordinary text; 2,000 cover something like 85–90 per cent of everyday conversation. After that each additional thousand adds only a percentage point or two.

The same curve holds for German, Spanish, Georgian, Korean — the numbers shift slightly, the shape does not. This is Zipf's law, and for a learner it says one thing: the first two thousand words are worth more than the next twenty thousand put together.

What "coverage" feels like

Eighty per cent coverage does not mean you understand 80 per cent of what people say. Meaning tends to sit in the rarer words: you can know every word in "the ___ was ___ by the ___" and still be lost.

But it does mean the sentence has a frame you can hold on to. You know where the verb is, what is being done to what, roughly whether it is good or bad. From that frame you can guess the missing word, or ask about it, or look it up — and looking up one word in a sentence you otherwise follow is how vocabulary grows on its own.

At 40 per cent coverage there is no frame. Every sentence is noise, and nothing sticks.

The wrong order is expensive

Most textbooks and phrasebooks start with topics: "At the airport", "Colours", "The family". Some of that is frequent — mother, water, good. Much of it is not. Luggage carousel, turquoise and nephew are perfectly real words that you will meet a few times a year, and every one of them takes as long to learn as because, which you will meet fifty times a day.

Learning by topic feels organised. Learning by frequency is what actually shortens the time until you can follow a conversation.

How Mullawo orders the deck

Every word in Mullawo's dictionary carries a frequency rank, drawn from large corpora of everyday English and applied to its translations in all 30 languages. New cards are dealt in that order: the first hundred you see are the words you will meet in the first hundred minutes of real speech.

Two things soften the pure ranking, on purpose:

• The free set is hand-picked, not the top 50 by rank. The most frequent words in any language are the, of, and — real, useful and hopeless as first flashcards. The opening deck is fifty concrete, frequent words: water, house, to eat, good.
• Homonyms are split. Train the noun and train the verb are two cards with two ranks, because they are two words in every language you are learning.

What to do with this

1. Trust the order. The urge to jump ahead to the words you want — the ones from your job, your hobby — is strong and mostly wrong for the first month. Add them as custom words if you like, but let the frequency deck run.
2. Do not skip the boring ones. Also, still, already, yet are the words that hold sentences together; they are dull as cards and priceless in speech.
3. Count coverage, not cards. A thousand frequent words is a milestone with a meaning: you can now follow the frame of most sentences. A thousand random words is just a number.