How many French words do you need?
It depends on the book, which is the thing the usual answer never says. Here is the answer for one shelf, counted: five fairy tales by Perrault and three stories by Daudet, 12,660 running words drawn from 1,610 distinct dictionary words. Every figure below is computed from that text when the page is asked for.
The answer, at five depths
| Dictionary words | Of every word printed | What that feels like |
|---|---|---|
| 32 | half | the articles, prepositions and pronouns — you can see the shape of a sentence and not what happens in it |
| 257 | 80% | one word in five unknown: decoding, not reading |
| 572 | 90% | one in ten: every sentence still costs a stop |
| 977 | 95% | one in twenty: readable with help to hand, and slow |
| 1,357 | 98% | one in fifty: the depth the reading research associates with reading unassisted |
So: 977 words for 95% of these eight real French tales, and 1,357 for 98%. The whole shelf, every word Perrault and Daudet use in them, is 1,610. A shelf this size is small — 50% of its words appear once — so the deeper figures sit close to its whole vocabulary, and a full-length novel asks for more.
The first 32 words are half of every page
The commonest dictionary words on this shelf, with the number of times each is printed:
| French | English | Times printed |
|---|---|---|
| le | the | 1017 |
| de | of, from | 569 |
| que | that, which, whom | 351 |
| et | and | 335 |
| être | to be | 300 |
| un | a, an | 244 |
| avoir | to have | 235 |
| son | his, her, its | 231 |
What counts as a word in French
A count of French words has to decide what to do with «l’enfant» and «qu’il». This one counts the glued word on its own and files it under its full form — «l’» under «le» — and files a verb with a pronoun hung on it, «dit-il», under the verb. It is the way a dictionary sees them, and it is why a French figure here is not directly comparable with a count that glued them together.
Coverage is not reading
Take 977 words — enough to cover 95% of the print — in the order that finishes whole sentences soonest, and of the shelf's 679 sentences that are more than a name or a cry, only 504 have nothing unknown left in them. At 1,357 words it is 624. The missing words are spread one or two to a sentence, which is why nobody crosses a line into reading; the number of sentences that come free simply grows.
Half the words appear once
812 of the 1,610 words on this shelf — 50% — are printed exactly once. They sit at the bottom of every frequency list, so no course prepares you for them, and looking each one up is what stops a reading. A glossed text is the answer: in the reading room every one of them carries its dictionary form and meaning where it stands.
Start with the shortest
«Le petit Chaperon rouge» is 683 words, free and entire, with no account. The word lists are ordered by how many whole sentences each word finishes, and the reading order prices every tale in words you do not yet have.