Vocabulary size

How many words do you need in a foreign language?

To read novels and newspapers without a dictionary, you need about 8,000–9,000 word families; for films and unscripted speech, about 6,000–7,000. These are Paul Nation's estimates for knowing 98% of the running words, the share research suggests most learners need to understand a text unaided. With about 3,000 families you already know roughly 95% of the words in films and TV programmes, enough to follow a good deal with some gaps.

By the Readlang Mobile team · Updated 28 September 2026

The figures depend on coverage

Coverage is the share of running words in a text that a reader already knows. The research on vocabulary size starts from a question about coverage: how many words can be unknown before understanding breaks down? Then it asks how many words a learner needs to know to reach that share in a given kind of text.

The best-known experiment is by Hu and Nation (2000). Sixty-six adult learners of English read a 673-word story in which some less frequent words had been replaced by invented ones, so the researchers could control exactly how much of the text was known. A reader counted as understanding the story adequately with at least 12 correct answers out of 14.

A 98% version of the story was never tested. The authors drew a line through these results and estimated that about 98% coverage is needed for most learners to understand fiction without help. That estimate is the basis of most of the figures on this page. It deserves some caution: a 2023 replication with 104 adult learners in Sri Lanka could not fully reproduce the original results (Kremmel et al., 2023).

An earlier study by Batia Laufer (1989) found that 95% coverage was the point at which significantly more learners reached a pass mark of 55% on a reading test, a much lower bar than comfortable reading. Laufer and Ravenhorst-Kalovski (2010) later treated the two figures as two thresholds: 95% as a minimum for reading with effort, and 98% as the optimum.

In everyday terms, 98% coverage means one unknown word in every 50; 95% means one in every 20.

How many word families for each goal

The ranges hide variation between texts. In the film study, 98% took 6,000 families for American films and 7,000 for British ones, and between 5,000 and 10,000 depending on genre. In spoken English, Nation found that 3,000 families plus proper nouns covered about 96% of recorded conversation. A study of a five-million-word corpus of everyday British speech found that the 2,000 most frequent word families covered a little under 95% of it (Adolphs & Schmitt, 2003).

The first few thousand words do most of the work

Words are not equally useful. Nation's (2006) counts show how quickly coverage rises and then flattens:

78–81%

of written text is covered by the 1,000 most frequent word families

81–84%

of spoken text is covered by the same 1,000 families

+8–9%

added by the second 1,000 families in written text (+5–6% in speech)

98.24%

of five novels is covered by 9,000 families plus proper nouns, against 87.83% for 2,000

Moving from 2,000 to 9,000 families adds only about ten percentage points of coverage in novels. But those points are the difference between meeting an unknown word about once in every eight words and once in every fifty, which is the difference between decoding a novel and reading it.

What counts as a word

All the figures above count word families. A word family is a base word with the inflected and derived forms a learner can understand without learning each one separately: watch, watches, watched, watching, and under wider definitions derived forms such as watcher (Bauer & Nation, 1993). Bauer and Nation arranged English prefixes and suffixes into seven levels, so the size of a family depends on the level chosen; Nation (2006) used level 6 families built from the British National Corpus.

Counting families matters because other counts give very different numbers. Our own counts of 117 books list distinct word forms, where watch and watched are two words. A novel that contains 6,000 different forms may use far fewer families, and in languages with many verb endings, such as Spanish or Italian, the gap is wider still.

How big is a native speaker's vocabulary?

Estimates depend on how words are counted, and the main studies count differently, so their numbers should not be set against each other directly.

• Goulden, Nation and Read (1990) estimated that well-educated adult native speakers of English know around 17,000 base words, which works out at two to three new words a day. Nation and Waring (1997) put a university graduate at around 20,000 word families, with native speakers adding roughly 1,000 families a year.
• Brysbaert, Stevens, Mandera and Keuleers (2016) estimated that the average 20-year-old native speaker of American English knows 42,000 lemmas (dictionary headwords), derived from 11,100 word families, with a range from 27,000 lemmas for the lowest 5% to 52,000 for the highest 5%.

Either way, the 8,000–9,000 families needed to read novels are well below a native speaker's vocabulary. Comfortable reading in a foreign language does not require native-like size.

How much reading does it take?

Nation (2014) worked out how much text a learner needs to read to meet each word of a frequency level about 12 times, a rough estimate of the repetitions needed to learn it. The answer grows quickly: just under 200,000 running words to meet the second 1,000 families often enough, and about 3 million for the ninth 1,000. At 120,000 words a novel, that is from about two to about twenty-five novels.

Spread over one level a year, reading at 150 words a minute, five days a week for forty weeks, that means about 7 minutes a day for the second 1,000 families, 33 minutes for the fifth, and 1 hour 40 minutes for the ninth. The calculation assumes texts in which no more than 2% of the words are unknown, which for most learners means graded or simplified books until their vocabulary is large.

How to measure your own vocabulary

For English, the most widely used measure is the Vocabulary Size Test by Nation and Beglar (2007). The original version has 140 multiple-choice items, ten from each 1,000-family level up to 14,000; your score times 100 estimates your size. Two newer 20,000-family versions have 100 items each, and the score is multiplied by 200. The tests are free to download from Paul Nation's resources page and can be taken online at my.vocabularysize.com. They measure only whether you recognise a word's meaning when reading, not whether you can use it.

For other languages there is no equally standard free test. The Council of Europe's CEFR self-assessment grid, linked from our first-week plan, is a reasonable starting point, followed by the practical check of counting unknown words on a page.

What this means for learners

• Learn the most frequent 2,000–3,000 families deliberately. They make up most of every text and every conversation, so time spent on them with a course or flashcards pays off everywhere.
• Beyond that, rely on volume. There are too many less frequent words, each too rare, to learn them all from lists. Reading and listening a lot at 95–98% coverage is how most of them are met often enough.
• Choose texts by coverage, not by ambition. One unknown word in fifty is comfortable; one in twenty is hard work. The two-minute check tells you where a book falls.
• Do not worry about names. Proper nouns made up about 1.5% of the words in the novels Nation analysed, and they are usually clear from context.

References

1. Adolphs, S., & Schmitt, N. (2003). Lexical coverage of spoken discourse. Applied Linguistics, 24(4), 425–438. doi:10.1093/applin/24.4.425
2. Bauer, L., & Nation, P. (1993). Word families. International Journal of Lexicography, 6(4), 253–279. doi:10.1093/ijl/6.4.253
3. Brysbaert, M., Stevens, M., Mandera, P., & Keuleers, E. (2016). How many words do we know? Practical estimates of vocabulary size dependent on word definition, the degree of language input and the participant's age. Frontiers in Psychology, 7, 1116. doi:10.3389/fpsyg.2016.01116
4. Goulden, R., Nation, P., & Read, J. (1990). How large can a receptive vocabulary be? Applied Linguistics, 11(4), 341–363. doi:10.1093/applin/11.4.341
5. Hu, M., & Nation, P. (2000). Unknown vocabulary density and reading comprehension. Reading in a Foreign Language, 13(1), 403–430. Open-access PDF
6. Kremmel, B., Indrarathne, B., Kormos, J., & Suzuki, S. (2023). Unknown vocabulary density and reading comprehension: Replicating Hu and Nation (2000). Language Learning, 73(4), 1127–1163. doi:10.1111/lang.12622
7. Laufer, B., & Ravenhorst-Kalovski, G. C. (2010). Lexical threshold revisited: Lexical text coverage, learners' vocabulary size and reading comprehension. Reading in a Foreign Language, 22(1), 15–30. Full text (ERIC)
8. Nation, I. S. P. (2006). How large a vocabulary is needed for reading and listening? Canadian Modern Language Review, 63(1), 59–82. doi:10.3138/cmlr.63.1.59
9. Nation, I. S. P., & Beglar, D. (2007). A vocabulary size test. The Language Teacher, 31(7), 9–13. Open-access PDF
10. Nation, P. (2014). How much input do you need to learn the most frequent 9,000 words? Reading in a Foreign Language, 26(2), 1–16. Full text (ERIC)
11. Nation, P., & Waring, R. (1997). Vocabulary size, text coverage and word lists. In N. Schmitt & M. McCarthy (Eds.), Vocabulary: Description, acquisition and pedagogy (pp. 6–19). Cambridge University Press.
12. Webb, S., & Rodgers, M. P. H. (2009a). The lexical coverage of movies. Applied Linguistics, 30(3), 407–427. doi:10.1093/applin/amp010
13. Webb, S., & Rodgers, M. P. H. (2009b). Vocabulary demands of television programs. Language Learning, 59(2), 335–366. doi:10.1111/j.1467-9922.2009.00509.x