Vocabulary size

How many words did Shakespeare know?

At least 66,500 words, counted as word forms, by the best-known estimate. His surviving works contain 884,647 running words and 31,534 different words, and from how often those words recur, the statisticians Bradley Efron and Ronald Thisted (1976) calculated that he knew at least 35,000 more that he never used in writing.

By the Readlang Mobile team · Updated 2 November 2026

How many different words are in Shakespeare's works?

About 29,000 to 31,500, depending on who counts and how. Both figures come from computer counts of the whole canon, the plays and poems, which add up to 884,647 running words. They differ because they group those words into "different words" in slightly different ways:

Most of these words are rare. In Efron and Thisted's table, 14,376 of the 31,534 words, nearly half, appear exactly once in the canon, and another 4,343 appear twice. Only 846 appear more than 100 times (ShakeWords data, from the same paper). That long tail of words used once is typical of large texts: in our counts of 117 books, too, about half of the different words in a typical English book appear only once.

Marvin Spevack's concordances, first published in nine volumes from 1968 (Spevack, 1968–1980), list every word in the works with every place it occurs. Counts of different words in the canon are made from concordances like these, and the figure of 29,066 comes from the one-volume Harvard Concordance.

Are these word forms, dictionary words or word families?

Word forms. In both counts, take, takes, taking and took are four different words. That makes the totals larger than a count of dictionary headwords (lemmas), and much larger than a count of word families, the unit used in most research on how many words learners need.

The unit matters whenever Shakespeare is compared with anyone else. The same modern vocabulary can be described as 11,100 word families, 42,000 lemmas or about 71,400 word forms (Brysbaert et al., 2016), as explained on how many words the average person knows. A Shakespeare figure in word forms should only be set against other figures in word forms, which is why the comparisons on this page all use them. Our own book counts are word forms too.

How do you estimate words he knew but never wrote?

By treating his works as a sample of his vocabulary, and asking how many new words a bigger sample would bring. Bradley Efron and Ronald Thisted (1976) borrowed the method from ecology, where it answers a similar question: how many species live in an area that no one has fully surveyed?

1. The butterfly problem. The statistician R. A. Fisher worked on butterflies that A. S. Corbet had collected in Malaya: some species had turned up many times, many only once or twice. Fisher, Corbet and Williams (1943) used the pattern of those counts to say how many new species further collecting would find. The rarer the species already seen, the more unseen ones there probably are.
2. Words as species. Each different word is a species, and each time Shakespeare uses it is one sighting. A word he used once could easily have gone unused, so many words he knew must have stayed below the line. The number of words seen once, twice, three times and so on tells you how many.
3. The Good–Toulmin estimate. I. J. Good and G. H. Toulmin (1956) gave a formula for the number of new species in a second sample: roughly the number of species seen once, minus those seen twice, plus those seen three times, and so on. Applied to Shakespeare, it predicts that a newly found set of works as long as the known canon would contain about 11,460 words that he never used in the canon (Efron & Thisted, 1976; summarised in Unseen species problem).
4. Going further. For samples much larger than the original, the simple formula breaks down. Efron and Thisted examined it alongside Fisher's parametric model, and extended it with linear programming, to see how many new words would turn up if he had gone on writing without limit. Their conclusion: the models are equivalent to supposing that Shakespeare knew at least 35,000 more words than the 31,534 he used.

31,534

different words in the surviving works

about 11,460

new words expected in another canon of the same length

at least 35,000

words he knew but did not use, by Efron and Thisted's estimate

at least 66,500

words in his vocabulary in total, in the same units

The estimate has limits that are easy to lose in retelling. It is a lower bound, not a central figure. It counts word forms, like the 31,534, so it is not a count of dictionary words. And it assumes that any further writing would have been of the same kind, plays and poems, so words that Shakespeare used only in conversation or business are outside its reach.

Was the method ever tested?

Yes, on a poem. In 1985 a short poem beginning "Shall I die?" was found in a manuscript and attributed to Shakespeare, and the attribution was disputed. Thisted and Efron (1987) used their model to test whether the poem's word usage fitted the canon. They concluded that, "on balance, the poem is found to fit previous Shakespearean usage reasonably well". That does not prove he wrote it, since another writer of the period could produce the same pattern, but it showed that the model makes predictions that can be checked against new text.

Did Shakespeare invent words?

Some, but far fewer than lists online claim, and no count is exact. The figures usually quoted come from the Oxford English Dictionary (OED), which gives the earliest quotation its editors found for each word. When that quotation is from Shakespeare, he is the first recorded user of the word, which is not the same as its inventor: someone may have said or written it earlier in a text the dictionary's readers never saw.

The tally has been falling. The first and second editions of the OED gave Shakespeare's works as the first evidence for around 2,000 words. As of June 2019, the revision for the third edition had reduced that to 1,467, as earlier uses turned up for many of them (Brewer). The revision is not finished, so the number can fall further.

Charlotte Brewer, who has studied how the OED was compiled, argues that Shakespeare's high count partly reflects how thoroughly the dictionary's first compilers recorded his language compared with that of his contemporaries, and that his one-off coinages were more likely to be recorded than similar ones by lesser-known writers (Brewer, 2012). The 1,467 words are also counted as dictionary entries, not word forms, so they cannot be set directly against the 31,534 different words in his works.

How does that compare with a modern adult?

In word forms, Shakespeare's estimated vocabulary is close to that of an educated young adult today, not far above it. Brysbaert, Stevens, Mandera and Keuleers (2016) estimated that the median 20-year-old native speaker of American English knows about 42,000 lemmas, which by their rough calculation is about 71,400 word forms. Efron and Thisted's lower bound for Shakespeare, about 66,500 word forms, is in the same range. Several differences cut both ways: Brysbaert's count excludes names, and the Shakespeare counts are of every form in the text; the English of around 1600 is not the English of today; and the two estimates come from completely different methods, one from a single author's writing, the other from a recognition test taken by over 200,000 people. The details of the modern estimates are on how many words the average person knows.

The words Shakespeare used can be compared with books, since our book counts are also word forms with names included:

Shakespeare's plays and poems together are longer than War and Peace, and they use more different forms than it does. Part of that comes from length: every new page brings some words not seen before, even in the same author. Part comes from range: history, comedy, tragedy and verse each bring their own vocabulary. How the number of different words grows with the length of a book, and how it differs between our 117 books, is on how many different words are in a book.

What does this mean for learners?

You do not need anything like Shakespeare's vocabulary to read in a foreign language. Paul Nation estimated that knowing 98% of the running words in a novel takes about 8,000–9,000 word families, with names counted as known (Nation, 2006), a far smaller unit and a far smaller number than the word forms above. What those thresholds mean and how they were measured is on how many words you need in a foreign language.

Shakespeare himself is a hard read for learners of English, and not only because of the size of his vocabulary. Nearly half of his different words appear only once, so a reader keeps meeting words that never come back, and many of his words have changed meaning or dropped out of use since. Our 117-book table includes Romeo and Juliet alongside novels in five languages, measured the same way.

References

1. Brewer, C. Shakespeare. Examining the OED, Hertford College, University of Oxford. Last updated 6 August 2025. oed.hertford.ox.ac.uk, checked 1 November 2026.
2. Brewer, C. (2012). Shakespeare, word-coining and the OED. In P. Holland (Ed.), Shakespeare Survey 65 (pp. 345–357). Cambridge University Press. doi:10.1017/SSO9781139170000.026
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. Efron, B., & Thisted, R. (1976). Estimating the number of unseen species: How many words did Shakespeare know? Biometrika, 63(3), 435–447. doi:10.1093/biomet/63.3.435
5. Fisher, R. A., Corbet, A. S., & Williams, C. B. (1943). The relation between the number of species and the number of individuals in a random sample of an animal population. Journal of Animal Ecology, 12(1), 42–58. doi:10.2307/1411
6. Good, I. J., & Toulmin, G. H. (1956). The number of new species, and the increase in population coverage, when a sample is increased. Biometrika, 43(1–2), 45–63. doi:10.1093/biomet/43.1-2.45
7. 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
8. Nevalainen, T. (1999). Early Modern English lexis and semantics. In R. Lass (Ed.), The Cambridge History of the English Language, Vol. III: 1476–1776. Cambridge University Press. doi:10.1017/CHOL9780521264761.006
9. ShakeWords: Shakespeare's word type frequencies, from Efron & Thisted (1976). Documentation of the R package vcdExtra. search.r-project.org, checked 1 November 2026.
10. Spevack, M. (1968–1980). A Complete and Systematic Concordance to the Works of Shakespeare (9 vols.). Hildesheim: Georg Olms. Folger Shakespeare Library catalogue
11. Spevack, M. (1973). The Harvard Concordance to Shakespeare. Cambridge, MA: Belknap Press of Harvard University Press.
12. Thisted, R., & Efron, B. (1987). Did Shakespeare write a newly-discovered poem? Biometrika, 74(3), 445–455. doi:10.1093/biomet/74.3.445
13. Unseen species problem. Wikipedia, section on Shakespeare's vocabulary. en.wikipedia.org, checked 1 November 2026.