Shakespeare's IQ: a 210 with no test, no case history

0 The short answer

William Shakespeare has no documented IQ: no intelligence test, examiner or date is attached to the figures of 160 to 210 that circulate. Catharine Cox's 1926 study of 301 eminent people printed no average rating above 200, and the index of her volume sends Shakespeare only to a table that, by her own account, also holds the names of eleven cases she discarded for lack of records. What has been measured is his text: Efron and Thisted counted 31,534 different word forms in the printed works. That is a count of words, which is not an IQ.

1 Where do the figures 160 and 210 come from?

Every figure for Shakespeare traces to a page that prints it without a test, an examiner or a date of measurement, and none traces to a document. The pages below were read on October 6, 2026. They disagree with each other on the number, and the ones that name any basis name his writing, his vocabulary or other web pages, never an assessment.

The Reader's Digest list and the Big Think list that the third page names both refused automated reading (HTTP 403) on October 6, 2026, so this page does not report what either prints. What the table does show is that the two ends of the circulating range, 160 and 210, are separated by 50 points, which is more than three standard deviations on the scale where the mean is 100 and the standard deviation is 15 (see why the standard deviation is 15).

One page that gives a reason cites vocabulary, a count of 30,000+ words used as if it were a score. Later sections trace such counts to their sources and show what they can and cannot support. Before that, the 1926 study that rated historical figures from biographical records needs a precise reading, because it is the likeliest place a reader would look for a Shakespeare rating.

2 Is Shakespeare in Catharine Cox's 1926 table?

Shakespeare is not in Cox's case histories and not in her ratings discussion: the index of her volume lists him once, under Table 12A, with no page for a case study. Cox's 1926 book, The Early Mental Traits of Three Hundred Geniuses, is the 1926 study that estimated IQs for 301 historical figures from biographical records (the method is explained on the page about how Cox rated Napoleon and in the audit of the highest IQ claims; it is not repeated here). Shakespeare was born in 1564, so he falls inside her window of people born after 1450. The question is what her volume prints for him.

The text that was read for this page is an OCR of the library copy, so the findings below are limited to what it contains. They are these:

1. The index entry. It reads "Shakespeare, William, Table 12A (facing 60)" and nothing else. Entries for people who have case histories continue with further page numbers, as the entry for Bacon does.
2. No case history. The contents list of the case histories names, among others, Jonson at page 291, Spenser at page 436, Bacon at page 501 and Milton at page 545. It names no Shakespeare. The body of the book names him as something other subjects read or are compared with, and once as an actor: in the case of Ben Jonson, Cox writes that Shakespeare performed in Jonson's revised first play (page 291).
3. Nine other names with the same entry. In the same index, Camoens, Condillac, Knox, LeSage, Loyola, Machiavelli, Pizarro, Rabelais and Titian also point only to Table 12A (facing 60).
4. Cox's account of the eleven. On page 32 she writes that the final group excludes "eleven cases for whom no records are available" and that their names are included for reference with those of Group A in Table 12A, opposite page 60. A footnote on page 38 gives the arithmetic: 282 cases in Group A, 19 in Group B and 11 members discarded from Group A for lack of data. On page 39 she repeats that eleven cases were discarded for lack of data and that an individual case history was prepared for each of the remaining 301.
5. Where the 301 sits. The group of 282 was drawn from a list of eminent names in which, Cox says, those born before 1450 were omitted and names beyond the 510th on the original list were stricken out (page 32).

Table 12A is a folded insert, and it is not present in the text read for this page. Therefore this page cannot print the row, if there is one. It can state two things with confidence. First, apart from the index entry, no case history, no rating discussion and no reliability coefficient for Shakespeare appears in the text read. Second, the pattern is exactly the one Cox describes for the eleven discarded cases: a name in the index that points only to Table 12A. The index search found ten such names, one fewer than her eleven, which can be an OCR gap or a difference in how the index was compiled. So the finding is stated as follows: Shakespeare was either one of the eleven cases without records or was listed in Table 12A for another reason that her text does not explain, and in neither case does the volume contain a rating that anyone can quote.

There is a second thing the book settles, and it matters for the number 210. In Chapter X Cox discusses the group at each rating level. She reports no cases at an average rating of 195 or 200 on the estimate to age 17, no cases at 195 on the estimate to age 26, and one case at 200 on the estimate to age 26, Goethe, whose early record she calls unusually complete (page 162; the case is on the page about Goethe). The highest average rating in that discussion is therefore 200. A 210 for Shakespeare cannot be one of the Cox averages for her rated cases. When Cox names the cases that in her view "probably rated at 200 IQ or even higher" (page 217), she lists Arnauld, Comte, Goethe, Grotius, Laplace, Leopardi, Michelangelo, Newton, Pascal, the younger Pitt, Sarpi, Schelling, Voltaire and Wolsey. Shakespeare is not on the list.

3 What would Cox have needed, and what does the record of Shakespeare's first 26 years hold?

A Cox rating is a judgment about a paper record of the first 17 and the first 26 years of a life, so it can only be as good as the record, and the record for Shakespeare's early years is mostly registers rather than performance. Cox's raters did not test anyone. Each subject received two ratings, each the average of three independent estimates: the AI rating, covering the record to the 17th birthday, and the AII rating, covering the record to the 26th (page 48). Each rating came with a reliability coefficient for the data behind it, and Cox reports a relation she describes as unexpected: the more reliable the data, the higher the estimate (page 52). Raters score what a record shows, and a thin record shows little.

Cox also warns against reading the numbers as measurements of persons. She compares John Stuart Mill, with an AI rating of 190, and John Milton, with 145, and writes that this should not be taken to mean that Mill's true score was 45 points higher, because each score registers the average rating of all individuals who do the things that are recorded of the given subject at a given age (page 48). That passage is the best single guide to every number in her volume, and it applies to any number a web page borrows from it.

The case histories Cox built show what she looked for. The headings in the case of Francis Bacon, for example, run through family standing, interests, education, school standing and progress, friends and associates, reading, evidences of precocity, and the development from 17 to 26 (page 501). Those are the items that decide a rating: what the child read, how quickly he advanced at school, what he produced early.

What do the pages read for this article hold for the same years of Shakespeare's life? The records are of a different kind:

• A baptism. According to Shakespeare Documented, a Folger resource, the parish register of Holy Trinity Church in Stratford-upon-Avon records on April 26, 1564 the baptism of William, son of John Shakespeare, and the register is held by the Shakespeare Birthplace Trust. The date of birth is not recorded. The page says it is traditionally given as April 23 and was probably a few days before the baptism.
• A school whose registers do not survive. The Shakespeare Birthplace Trust says, in a podcast transcript with scholars, that no records or registers from the school at the time survive, and that a boy in his position would have been entitled to a free place at the grammar school. The same transcript describes a Latin-centered curriculum taught by Oxford graduates. That is a description of what the school offered, not a record of how he did there.
• A marriage and children. The Trust's biography page states that he married Anne Hathaway at the age of eighteen and that twins were baptised in 1585.

None of these is a record of reading, school standing or early production, the items Cox's raters scored. That is a statement about the sources read for this page, not a claim that nothing else exists. It does explain why a historiometric rating of the early years would be hard to anchor, and Cox's own rule was to leave out the cases for which no records are available rather than score them.

4 How do the ratings of Shakespeare's fellow writers show what a thin record does?

Cox's ratings for the writers nearest Shakespeare in time and trade rise and fall with the reliability of their records, not with any difference anyone measured between the men. The table sets out what her case pages print for Edmund Spenser, Ben Jonson, Francis Bacon and John Milton. It is included because it shows the mechanism a Shakespeare entry would have been subject to.

Spenser is the clearest case. Cox writes that the two ratings in his group based on the least reliable data, .11 each, are those of Erasmus and Spenser, and that Spenser's is low because of a question about the true authorship of certain translations sometimes attributed to the young poet in his 16th year. If the supposed early achievement were proved genuine, she says, he would be rated near 150, and the uncertainty of a judgment that depends on uncorroborated evidence led the raters to regress toward 100 (page 112). That is the method working as designed: when the evidence is contested, the rating shrinks toward the middle. The ratings track what is recorded, and they say little else.

The four rows should not be read as a ranking of the men. Cox herself warns, as above, against reading rating differences as differences in true intelligence. Nor should anyone fill the gap in Table 12A by splitting the difference between neighbors, because the neighbors differ by their records, not by their minds. Any figure inferred for Shakespeare from the Milton and Bacon rows would be invented, and this page does not offer one.

5 What does the documentary record of Shakespeare's life actually hold?

The documents that survive for Shakespeare establish who he was, when he lived and what was printed under his name, and they say almost nothing about how he thought. They are worth stating soberly, because the figures on the lists lean on a sense that something documented must lie behind them.

• The will. Shakespeare's will is dated March 25, 1616, is held at The National Archives at Kew, and carries three signatures, one on each of its three sheets. The Folger page reports that it mentions no books and no plays. Its bequests include money to his daughter Judith, property to his daughter Susanna Hall, a bequest of the second best bed to his wife Anne Hathaway, and 26 shillings and 8 pence each to his theater colleagues Burbage, Heminges and Condell.
• The First Folio. The 1623 collected edition, published in London by Isaac Jaggard and Ed. Blount, is described on the Folger's page as the first edition and sole source for half of the plays, among them Julius Caesar, Macbeth and The Tempest. That is a fact about the survival of texts, and it is why the plays can be counted at all.
• The canon itself. The Folger's frequently asked questions give the complete works as 884,647 words and 118,406 lines, add that counts may differ with the criteria of what is included, and call 38 the generally accepted number of plays. These numbers are the raw material of every vocabulary study in the next sections.
• An actor, in another man's case. Cox's volume, in the case of Ben Jonson, writes that Shakespeare performed in Jonson's revised first play (page 291). It is the only place the book gives him a role.

The authorship dispute is outside the scope of this page, and the page takes no position beyond the documents above. Every figure traced here attaches to the works printed under his name, and nothing in the argument changes if a reader attributes those works to someone else, because no intelligence test was ever given to the author of these works, whoever the reader takes him to be.

6 How many words are in Shakespeare's works, and who counted them?

The total of 884,647 words is a count of running text, the 31,534 different forms are a count of types, and ranges such as "17,000 to 29,000" mix counts made with different rules. A count of words in a printed text is the one quantity in this whole story that can be measured exactly, as long as the rules for what counts are stated.

The Folger gives the running words of the complete works as 884,647. Efron and Thisted, in their 1976 Biometrika paper, report in the abstract that Shakespeare wrote 31,534 different words, of which 14,376 appear only once and 4,343 twice. The R documentation for the data set built from that table, vcdExtra's ShakeWords, gives the same totals and adds that 846 words appear more than 100 times. Our arithmetic on these published numbers: words that appear exactly once make up about 46 percent of the 31,534 different forms (14,376 divided by 31,534). That large share of one-time words is what the 1976 method works from, as the next section explains.

What counts as "a word" changes the answer. James Jackson, writing in Significance magazine on December 21, 2023, notes that the Efron and Thisted analysis included proper nouns and treated words as case-sensitive, which can count one word twice. Charlotte Brewer, in her 2012 chapter in Shakespeare Survey, adds a third layer. Marvin Spevack's concordance, she writes, gives a total of 30,000 words used by Shakespeare, but that count includes different forms of the same lexeme, and his entire recorded word stock comes to "something under 20,000 lexemes". A range such as 17,000 to 29,000 therefore mixes two units, forms and base words, and neither is a measure of what he knew. Spevack's own volumes were not opened for this page, so the exact pair 17,000 and 29,000 is not attributed to him here.

7 How many words did Shakespeare know? What Efron and Thisted estimated in 1976

Efron and Thisted estimated a floor, not a vocabulary: their models are equivalent to supposing that Shakespeare knew at least 35,000 words that he never used in the printed works. The abstract of their paper, "Estimating the number of unseen species: How many words did Shakespeare know?", says that they used both a parametric empirical Bayes approach and a nonparametric approach with linear programming, and that the two arrive at that conclusion. The title borrows a problem from ecology: if a naturalist has trapped a sample of animals and counted how many species appeared once, twice and so on, how many species were never trapped? Here the species are different words and the sample is the canon.

Our arithmetic on the published figures: 31,534 different forms used, plus at least 35,000 unused, gives at least about 66,500 different word forms, in the unit the paper counts. The paper does not state that total in the abstract, so this page does not attribute it to the authors.

Four limits apply, and each comes from sources that were read for this page.

1. The model is an assumption about how text is produced. Peter McCullagh, in a 2007 introduction to the paper, explains that the method treats an author as having a personal vocabulary from which text is generated by sampling, and that stationarity is needed for prediction and inference. He adds that stylistic and literary arguments do not support random sampling, yet the word-frequency data are consistent with it, and the fit is "extremely good".
2. The simplest version has no upper limit. McCullagh notes that in the model's simplest form the estimate for Shakespeare comes out unbounded, and that this is not unreasonable, since his vocabulary presumably grew over time. That fits a result stated as "at least", not as a point estimate.
3. It counts forms, not base words. The unit is the one in the data, including inflected forms and proper nouns (see the previous section), so the estimate cannot be set beside a count of lexemes.
4. It estimates knowledge from production. The method infers how many words an author could have used from the words he did use. It says nothing about how well he understood them, how fast he learned them, or what else he could do.

A second paper by the same authors tested the method on a new text. In Thisted and Efron's 1987 paper in Biometrika, the abstract reports that they examined word usage in a previously unknown nine-stanza poem attributed to Shakespeare, with comparison poems by Jonson, Marlowe and Donne and four confirmed Shakespeare works, and that "on balance" the poem fits previous Shakespearean usage reasonably well. McCullagh adds the numbers: the poem, found in the Bodleian Library, has 429 words, nine of them not previously seen in Shakespeare's writings, and the model had predicted 6.97. That is a test of authorship by word usage. It is not a test of intelligence, and nothing in either paper claims otherwise.

How does the floor compare with ordinary vocabularies? Brysbaert and colleagues, in a 2016 study in Frontiers in Psychology, estimated from the literature and a large crowdsourcing experiment that an average 20-year-old native speaker of American English knows 42,000 lemmas, with a range from 27,000 for the lowest 5 percent to 52,000 for the highest 5 percent, and that between ages 20 and 60 people learn about one new lemma every two days. A lemma is an uninflected word from which inflected words are derived, and the authors note that knowing a word can be as shallow as knowing that it exists. The two numbers cannot be compared directly: one counts word forms in a text and the other counts lemmas that people recognize. No study found for this page puts them in the same unit, so a sentence such as "he knew more words than ordinary people" has not been shown by either source.

8 Did Shakespeare coin 1,700 words?

The figure of about 1,700 is a count of words whose first recorded use in the Oxford English Dictionary is in his works, adjusted by one scholar, and it is neither a count of invented words nor a settled number. Charlotte Brewer's 2012 chapter, Shakespeare, Word-Coining and the OED, is the clearest account of how the figure was made and of its limits. The chapter is also dated: it works from OED data as of 2011.

Brewer reports the starting point. David Crystal, searching the second edition of the OED on CD-ROM, found Shakespeare recorded as the first person to use over 2,000 lexemes, 2,079 to be exact, which narrowed to 2,035 after malapropisms and nonsense words were excluded (Crystal, The Stories of English, 2004, as cited by Brewer). Brewer notes that this is about one in ten of his lexemes, a proportion that no other post-medieval writer in the OED matches. Her comparison of absolute counts is just as stark: Philemon Holland has around 1,000 first citations, Thomas Browne around 830, Henry More around 690, Milton about 650, Carlyle about 565, Ben Jonson about 560, Coleridge about 520 and Sidney about 400.

Crystal then adjusted for the chance that Shakespeare was simply the first writer to be quoted. He compared the 2,035 words by how soon each was recorded again from another source. Brewer reports his result: some 640 appeared again within twenty-five years, and about 1,100 only after twenty-five years. Crystal's sum is 300 words recorded in the OED only in Shakespeare, plus 1,100 or so recorded again only after a long gap, plus about 320, half of the 640, that he may have made up and that his contemporaries adopted quickly. That gives around 1,700.

Brewer then lists why the figure cannot be read as a count of inventions:

• The dictionary favored him. She writes that the OED's preference for citing Shakespeare, together with the reduced number of sixteenth and seventeenth century texts available while the first edition was compiled, will have led the dictionary to favor him over his contemporaries. Fresh reading finds earlier uses: newsmonger, merited, protractive, reprieve, repure, roguery and sportive occur earlier in Thomas Nashe, ratcatcher was pushed back to 1565, and ransomless to a text of 1425. A 1980 study by Jürgen Schäfer, Documentation in the OED, is among those she cites for the doubt.
• Print is not speech. The dictionary confines itself to printed sources, so a word may have been in spoken use long before it appeared in a play.
• Words that did not survive. Words such as conceptious look like inventions, but she asks whether a word that never achieved general currency can be said to have contributed to the language, and notes that such words may have entered the dictionary only because he used them.
• The dictionary is not a frequency corpus. Crystal's adjustment depends on the OED recording quotations consistently, but the dictionary sometimes printed two quotations from one decade and sometimes left gaps of decades, and its official policy was one quotation per century.
• The count can also be too low. Because the older search tools identified only one quotation per entry as the first, a writer who was first to use one sense of a word among several was not counted.

The most informative evidence is the revision. In the sample of entries from M to rotness that had been revised by November 2010, OED2 recorded Shakespeare as first user of the head-word in 406 entries and OED3 in 263, so earlier usages had been found for about one third. Brewer's projection, which she conditions on the sample being a trustworthy predictor, is that his total of 2,000 or so will drop to about 1,200, still unmatched by any other post-medieval writer recorded in the dictionary. In two subsamples of 50 first quotations each, 17 of 50 (34 percent) and 26 of 50 (52 percent) had been antedated. She also observes that the dictionary tells us he has just over one percent of its total number of quotations. Later revisions were not checked for this page, so the projected 1,200 should be read as of 2011 to 2012.

What does the coinage count measure? It measures the dictionary's record and the record's gaps. It is an important finding about language, and it makes no claim on intelligence: no study in this chain converts a count of first citations into a score on any scale.

9 Every number in one table: what each figure is, and whether it is an IQ

Of the eight figures that attach to Shakespeare in this story, none is a measured IQ, and only the word counts come from a procedure that anyone can repeat. The table sets each beside its origin so a reader can see which numbers are counts, which are projections and which are guesses.

The same pattern appears in the audits of other figures whose numbers circulate without a test. The page on Charles Darwin takes up the Cox rating of 135 attached to his name, and the page on Leonardo da Vinci weighs the figures of 180 and 220 attached to his. The general lesson, set out in the audit of celebrities with the highest IQ, is that a precise number is not evidence of a measurement.

10 Is a large vocabulary the same thing as a high IQ?

A large vocabulary is evidence of crystallized knowledge built up through years of exposure to language, and an IQ is a summary of several abilities of which that knowledge is only one. Raymond Cattell, in his 1963 paper in the Journal of Educational Psychology, is the standard reference for the distinction between fluid and crystallized intelligence. In the current CHC framework the second is called Gc, and it sits beside fluid reasoning (Gf), quantitative knowledge (Gq), visual processing (Gv), working memory (Gwm) and processing speed (Gs). The page on fluid versus crystallized intelligence covers the distinction, and the CHC model and what IQ measures place both within the larger structure.

Three features of the vocabulary evidence in this story keep it from working as an IQ.

1. It is production, not recognition. Efron and Thisted's counts are of words he put on the page. A vocabulary subtest samples words a person recognizes or defines. ACIS has such subtests in its verbal comprehension index, namely Vocabulary, Synonyms, Antonyms, Information and Similarities, and the page on the verbal IQ test explains what that score does and does not capture. Counting the words an author used is not an administration of any of them.
2. It depends on age and exposure. Brysbaert and colleagues report that people learn about one new lemma every two days between ages 20 and 60, and that 20-year-olds range from 27,000 lemmas at the low end to 52,000 at the high end. A vocabulary count is therefore partly a count of years and of reading. Anyone who wants to know whether such knowledge can be raised will find the evidence on what training can move.
3. It leaves out the other domains. A printed text contains no direct measure of fluid reasoning, working memory span, spatial processing, speed on a timed task or quantitative skill. The plays show craft in plotting and in language, but craft is a record of achievement, and an achievement record is not a score on a standardized task. The relation of creative work to IQ is covered on the page on IQ and creativity.

The practical point is simple. If the question is how large a vocabulary a writer used, Efron and Thisted answer it with a count. If the question is how that writer would have scored on a modern test, no document answers it, and the count does not stand in.

11 Why can no one assign Shakespeare an IQ today?

No one can assign Shakespeare an IQ because an IQ is a position within a norm group on a standardized task, and he never sat a task and no norm group of his time exists. The score on the scale with a mean of 100 and a standard deviation of 15 says how far a person stands from the average of a defined sample of living people, as explained on the page about how IQ scores are normed and the one on why the standard deviation is 15. The scale was built for people tested under standard conditions, and the page on the history of IQ testing shows how recent that practice is compared with the sixteenth century.

Three things stand in the way, and each has been documented above.

1. No task was ever given. The only quantitative studies of his work count words. The 1926 ratings, where they exist for others, score a paper record, and Cox states that each score registers what is recorded of the person at a given age (page 48).
2. The record for him is not the kind Cox scored. The registers, the will and the First Folio establish a life and a body of work. They do not contain the early reading, school standing and precocity that her raters needed, and the sources read for this page say that the school registers do not survive.
3. The number is outside what a norm sample can support. A score of 210 is (210 minus 100) divided by 15, or about 7.3 standard deviations above the mean, by our arithmetic. Norm samples of any realistic size contain no one that far from the mean, so a figure of that size cannot be tied to a percentile by evidence, and the page on how rare a given IQ is shows what such rarity would imply. The question of how rare the circulating figures would be if they were real is therefore answerable only as arithmetic, not as measurement.

A fair statement of what can be said is short. Shakespeare left a canon of 884,647 words that scholars have counted, parsed and compared. His documented life is a baptism, a marriage, a theater career, a will and a posthumous collected edition. The word counts show an author who used 31,534 different forms and who is modeled as having known at least 35,000 more. None of this supports a score, and the more careful pages on the lists say as much. The audit of the highest IQ ever recorded and the page on genius and IQ explain why such lists keep reappearing, and the page on Isaac Newton and the one on Albert Einstein trace two other figures that circulate without a test.

12 What does this mean for your own score?

A score is only as meaningful as the instrument, the norm group and the documentation behind it, and the standards for testing say the same about every score, including yours. The Standards for Educational and Psychological Testing, published by AERA, APA and NCME in 2014, ask for a clear articulation of each intended test score interpretation for a specified use and for validity evidence in support of each intended interpretation (Standard 1.0). If validity for a likely interpretation has not been evaluated, they ask that this fact be made clear and that users be strongly cautioned against unsupported interpretations (Standard 1.3), and if a score is interpreted in a way that has not been validated, they place on the user the job of justifying the new interpretation (Standard 1.4). A 210 attached to a name is an interpretation with no instrument, no norm and no validation behind it, and it fails the first test before the others arise.

The APA Guidelines for Psychological Assessment and Evaluation, approved by the APA Council of Representatives in March 2020, point the same way. Guideline 6 asks psychologists to select assessment tools that demonstrate sufficient validity evidence for their uses, sufficient score reliability and sound psychometric properties, and measures that are fair and appropriate for the purpose, population and setting. Guideline 7 asks for multiple sources of relevant and reliable information, and Guideline 8 asks for attention to the appropriate normative comparison, situational influences and standardized administration. Group statistics, vocabulary counts and biographies are not individual scores, and nothing here is a clinical statement about anyone, including the reader.

For a reader who wants a number of their own, the useful comparison is with a normed instrument rather than with a name from a list. ACIS is an online, self-administered assessment with 20 subtests in six domains, reported as a Full Scale IQ and six primary indices (verbal comprehension, fluid reasoning, quantitative reasoning, visual spatial, working memory and processing speed). Scores are on the scale with a mean of 100 and a standard deviation of 15, scaled subtest scores run from 1 to 19, and the report gives percentiles and a 95 percent confidence interval. Adult norms cover ages 16 to 90. Prices were read on October 6, 2026 and can change: Quick is 15 dollars for 6 subtests in 3 domains, Optimized is 30 dollars for 13 subtests in 5 domains, and Full Scale is 50 dollars for all 20 subtests, each a one time payment with no subscription. There is a free trial with no card and a 5 day quality guarantee, and each form can be completed over 30 days with breaks allowed.

The limits matter as much as the features. ACIS is online and unsupervised, it is not a clinical or diagnostic instrument, it is not for hiring, school accommodations or admission to high IQ societies, and it is available in English only. The technical manual documents the instrument. To see how a score converts to a standing, the page on IQ score versus percentile explains why the two cannot be averaged, and the pages on the best online IQ tests and on IQ tests for adults compare the options. A fair self-comparison is to your own earlier results or to a norm group, never to a historical figure, because the historical figures have no comparable number to offer.

13 Sources Behind This Page

Every figure above is traceable to one of the following, and each is linked at the point where it is used. Web pages were read on October 6, 2026. Figures marked as our arithmetic are calculations on published numbers and are not attributed to any author.

• Cox C M. The Early Mental Traits of Three Hundred Geniuses. Genetic Studies of Genius, volume 2, edited by L M Terman, Stanford University Press, 1926. Pages used: 32, 38, 39, 48, 52, 112, 156 to 162, 217, 291, 436, 501 and 545, the contents list and the index. Read in an OCR text of a library copy. Table 12A, a folded insert facing page 60, is not in that text and was not seen.
• Efron B and Thisted R. Estimating the number of unseen species: How many words did Shakespeare know? Biometrika, volume 63, issue 3, 1976, pages 435 to 447. The abstract was read on the Oxford Academic page; the full text was not.
• Thisted R and Efron B. Did Shakespeare write a newly-discovered poem? Biometrika, volume 74, issue 3, 1987, pages 445 to 455. The abstract was read on the Oxford Academic page; the full text was not.
• McCullagh P. Introduction to a paper by Efron and Thisted, "Estimating the number of unseen species: How many words did Shakespeare know?" University of Chicago, June 2007, three pages.
• Brewer C. Shakespeare, word-coining and the OED. Shakespeare Survey, Cambridge University Press, 2012, pages 345 to 357 (Crossref lists no volume). Read in the PDF hosted by the Examining the OED project at Oxford.
• Brysbaert M, Stevens M, Mandera P and Keuleers E. 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, volume 7, article 1116, July 29, 2016. The abstract and the definitions were read in the open access version on PubMed Central.
• Cattell R B. Theory of fluid and crystallized intelligence: A critical experiment. Journal of Educational Psychology, volume 54, issue 1, 1963, pages 1 to 22. Only the Crossref record was checked, and the paper is cited only for the distinction named in its title.
• Folger Shakespeare Library. Shakespeare Documented: parish register entry recording William Shakespeare's baptism, last will and testament and the First Folio. Pages read on October 6, 2026.
• Folger Shakespeare Library. Frequently asked questions about Shakespeare's works. Page read on October 6, 2026.
• Shakespeare Birthplace Trust. Was Shakespeare educated? Let's Talk Shakespeare podcast transcript. Page read on October 6, 2026.
• Shakespeare Birthplace Trust. William Shakespeare biography. Shakespedia. Page read on October 6, 2026.
• Jackson J. Words, words, words: How big was Shakespeare's vocabulary? Significance magazine, December 21, 2023.
• vcdExtra package documentation. ShakeWords: Shakespeare's word type frequencies. R documentation, page read on October 6, 2026. It cites Efron and Thisted (1976) as its source.
• American Educational Research Association, American Psychological Association and National Council on Measurement in Education. Standards for Educational and Psychological Testing. AERA, 2014. Standards 1.0, 1.3 and 1.4 were read in the open access PDF.
• American Psychological Association. APA Guidelines for Psychological Assessment and Evaluation. Approved by the APA Council of Representatives, March 2020. Guidelines 6, 7 and 8 were read in the PDF.
• Pages that print a figure for Shakespeare, read on October 6, 2026: geniuses.club, celebrityiqs.com, check-iq.org and lifescoretest.com.

14 Frequently Asked Questions