Language through Literature

How much Russian do you need to read Chekhov?

Ask this question anywhere and you get folklore: "a few thousand words", "just start", "Chekhov is easy". We can do better than folklore, because we ship the texts. What follows is measured on the fifteen classics in our library — Chekhov mostly, with Gogol, Dostoevsky and Turgenev — 68,172 running words of real, unadapted Russian prose in 5,075 sentences.

What the texts actually contain

Those 68,172 running words reduce to 7,746 distinct dictionary words (lemmas — every case and conjugation counted as one word), plus 273 proper names, which cost a reader nothing. The first uncomfortable fact lives inside that number: 3,824 of those lemmas — 49.4% — occur exactly once in the whole corpus. Half of Chekhov's vocabulary is words he uses one single time. You will never "finish" learning this list, and you do not need to: those words are what glosses are for.

The threshold that matters

The research question is not "how many words exist" but "what fraction of a page must you know for the rest to be guessable". Hu and Nation's much-cited study put the threshold for adequate unassisted comprehension at 95–98% coverage of the running words — at 95% some readers manage, at 98% most do (Hu & Nation, Reading in a Foreign Language, 2000). Below that, a page stops being a text with gaps and becomes a gap with text in it.

So the honest form of the question is: how many words buy you 95–98% of a real page? Our reader orders its vocabulary not by raw frequency but by how many whole sentences each next word completes — a 7,000-word ladder computed over the corpus. Here is exactly what climbing it buys, with names counted as known:

Read the middle column against the threshold. Two thousand words — a solid A2–B1 vocabulary — buys 82.8% coverage: still short of 95%, which is precisely why "just read Chekhov, he's simple" fails beginners. Coverage crosses 95% around five thousand words and reaches 98.9% at the top of the ladder. The full 95–98% band, the one the research says you need, sits at roughly five to seven thousand words — a B2 vocabulary, honestly earned.

The right-hand column is the consolation. A sentence with one unknown word is not a sentence you can read, so we also count sentences with no unknown words — and that number grows fast: at two thousand words more than half the corpus's sentences are already completely yours, and at the top of the ladder 95.6% are.

What to do with a vocabulary of hundreds, not thousands

Not wait — read with support, and read the short ones. The gap between your five hundred words and Chekhov's seven and a half thousand is exactly the gap a glossed reader closes: every word explained where it stands (15,261 word-forms glossed by hand in ours), the English one tap away, so the 30% you don't yet hold costs a glance instead of a dictionary. These are the shortest complete works on our shelf, all free, no account:

«Толстый и тонкий» is 511 words long. At that length you can afford a text where you know 70% of the words, because the whole undertaking is twenty minutes — and finishing a real Chekhov story at A2 does more for the next month of your Russian than any graded reader.

So: how much Russian do you need to read Chekhov? To read him unassisted, five to seven thousand words. To read him tonight, about five hundred and a good gloss.