VERBAVIA

How many Esperanto words do you know?

Esperanto · a two-minute vocabulary test · free, no sign-up

You will see 36 words, one at a time — about two minutes. Tap I know it only if you know what the word means; otherwise tap No. 6 of them are made up to look like Esperanto, so don't guess: a yes to a word that does not exist lowers your score.

On a keyboard: Y or → for yes, N or ← for no, Backspace to go back one.

This test runs in your browser and needs JavaScript switched on. Nothing is sent anywhere and nothing is stored.

Learn the words you are missing, in order

Start the free Esperanto course

The course deals this same list from the top down, one lesson a day, with the grammar that holds the words together. Or read the list itself: the most common Esperanto words, with translations.

About the Esperanto test

The Esperanto list is 7,400 words, and the test measures up to that. Esperanto builds words from roots and affixes, and if you can work out what a word means from its parts, that counts as knowing it — it is how the language is meant to be read. The made-up words are invented roots with ordinary endings and none of the affixes, so they cannot be decoded that way. The coverage figure comes from a small corpus of Esperanto subtitles, 403,882 words, and should be read as a rough guide.

How the number is worked out

The test draws on the top 7,400 entries of the course's Esperanto frequency list, cut into 5 bands. Names, conjugated forms glossed as forms of another word, and grammatical particles are set aside, because none of them is a word you could be asked about on its own; what is left is 7,318 words, and that is the most this test can measure.

BandList ranksWords it stands for
11–1,000964
21,001–2,000971
32,001–3,5001,497
43,501–5,0001,495
55,001–7,4002,391

Each attempt shows 6 words chosen at random from every band, and 6 made-up words, all shuffled together. Your yes-rate on each band is corrected by your yes-rate on the made-up words — the share of the words you do not know that you would say yes to anyway — using the standard correction for guessing: known share = (yes-rate on the band − yes-rate on made-up words) ÷ (1 − yes-rate on made-up words), never below zero. Each band's corrected share is multiplied by the number of words it stands for, and the bands are added up. If more than four in ten made-up words get a yes, no number is given, because there is then no telling knowledge from guessing.

With 6 words a band the test is quick, and that makes it an estimate, not a census: two attempts can differ by a few hundred words on a short list and by more on the wide top bands of a long one, so take it twice and split the difference if you want a steadier figure. It measures recognition — whether you know a word when you see it — which always runs ahead of the words you can produce yourself.

The second number is how much of ordinary Esperanto those words account for, and it is measured rather than assumed: the OpenSubtitles frequency counts — 403,882 words of film and television subtitles — say what share of all the words spoken there the most frequent 100, 1,000 or 10,000 make up. Your known share of each band is applied to that band's slice of the curve. Two cautions. The subtitle counts are of word forms, so hablo and habla count separately, while this test counts dictionary words; knowing a word usually means knowing several of its forms, so if anything the true figure is a little higher. And coverage is not comprehension: knowing nine words in ten still leaves a word you do not know in most sentences, and that is often the word that matters.

There is no ranking against other learners here, because we do not have that data and will not invent it. How much of real Esperanto the first 100, 1,000 and 10,000 words cover is measured in how many words do you need to speak Esperanto.

Tests in other languages