Nordic Margins
A large Tudor manor house of brick and timber with lit windows, reflected in a moat among old trees under a crescent moon

Identifying an Author: the Earl of Derby's Letters and Shakespeare

As a schoolboy Nordling thought the authorship question could be settled by comparing styles. Decades later he tried it with five letters signed Will Derby. This essay sets out his test and what it can and cannot bear.

Shakespeare and the authorship question 6 min read

Attributing anonymous texts by style is an old and respectable discipline. In 1962 the Swedish linguist Alvar Ellegård published Who Was Junius?, a statistical study of the anonymous letters that attacked George III's ministers in the 1760s and 1770s, and concluded that their author was Sir Philip Francis. Nordling admired the book and adopted its spirit: count features that writers use unconsciously, compare them with a control group, and see who fits.

He began by summarizing where his earlier essays had led. The description of the author drawn from the vocabulary of the plays, the proposed author of the German Hamlet, the model for the hesitant prince and the subject of the Stanley epitaphs all pointed, in his view, to William Stanley, 6th Earl of Derby. Treating "Shakespeare" as Derby's pen name would answer several questions at once, he argued: why the Stanley family had Shakespearean epitaphs, what became of the comedies Derby was reported writing in 1599, and why A Midsummer Night's Dream has often been linked with a wedding, since one of the occasions scholars have proposed for it is Derby's own marriage in January 1595.

Five letters signed Will Derby

The material for the test was five private letters signed "Will Derby," written between 1595 and 1607 and reproduced in facsimile by A. W. Titherley in Shakespeare's Identity (1952). Together they contain about 920 words. For the Stratford man there is nothing comparable, only six signatures.

Nordling picked out 30 turns of phrase in the letters that he judged not commonplace: expressions such as "to see service" for military duty, "to make bold," "to think meet," "to prefer a suit," "to ride post" and "at jar" for being in discord. Every one of them, he found, occurs somewhere in Shakespeare. For several, the Oxford English Dictionary cites Shakespeare as the earliest or one of the earliest users. Edmund Spenser, whose surviving verse is a large body of text, used only seven of the 30.

Then he applied arithmetic. If each phrase were used by half of all educated writers, the chance that a single writer other than Derby used all 30 would be about one in a billion (two to the power of 30). He compared this with fingerprints, where a dozen matching details are traditionally enough for a court.

Frequencies, rare words and Cytty

He added two frequency checks on Derby's longest letter, dated 10 July 1596 and nearly 400 words long. The 43 most common words in Shakespeare make up about 40 percent of his text, and about 39.5 percent of the letter. Words used fewer than 20 times in the whole canon make up about ten percent of Shakespeare's text and about 9.8 percent of the letter.

The most striking evidence was spelling. Nordling compared the letters with the three-page addition to Sir Thomas More in the hand known as Hand D, which many scholars believe is Shakespeare's own. In both he found -nes for -ness, -cion for -tion, howse, yf and yt, and the same abbreviations. Above all, he found the word city spelled Cytty in both, a form he said he had found in no other writer. He measured letter heights, spacing and slope and found no contradiction, while conceding that a layman's measurements prove little. He then compared spellings on the two Stanley epitaphs and in a poem preserved in what Titherley described as a manuscript in Derby's hand, a version of "When as thine eye hath chose the dame," which was printed in The Passionate Pilgrim (1599). Nordling called it "The Knack."

A control group of contemporaries

To show that the method discriminates, he applied it to letters by other men. Of 19 characteristic expressions he found in two letters by Edward de Vere, Earl of Oxford, five appear in Shakespeare. Of 18 in letters by Francis Bacon, five. Of 18 in letters by Lord Burghley, seven. Of 14 in the will of William Shakespeare of Stratford, two. Only Derby scored 30 out of 30.

Where the statistics are fragile

The test is clear and easy to repeat, which is to its credit. Its weaknesses are also clear.

The selection is not blind. Nordling chose which expressions in Derby's letters counted as distinctive, knowing he was looking for matches in a canon of almost 900,000 words. A canon that large contains a very high proportion of the idioms of its time, so a short letter by many educated writers might score well if the phrases were chosen the same way. The control group was not treated identically: the expressions taken from Oxford, Bacon and Burghley were defined by different criteria and in different numbers.

The probability figure assumes that each phrase is an independent coin toss used by half of all writers. Phrases like "to think meet" or "to make bold" were common in the formal correspondence of the period, so the true odds are unknown and probably far less dramatic. Modern computational stylometry generally treats a few hundred words as too short a sample for reliable attribution, and it compares function-word profiles over large texts, not hand-picked phrases.

The spelling argument depends on readings of manuscripts that specialists would want to check, and on the claim that a form like Cytty is unique, which requires a search of far more Elizabethan manuscripts than any single researcher could make. The attribution of Hand D to Shakespeare is itself an inference, though a widely accepted one.

None of this shows that Nordling was wrong, only that the test cannot carry the certainty he claimed for it. Derby's surviving writing is too small for modern methods, and no specialist study has taken up the comparison. Nordling held that the method could have been used as early as 1918, when Lefranc proposed Derby, and wondered whether anyone who tried it had been ignored. His account of why the profession might ignore such work follows in Credence and science. His lists of shared phrases are summarized in the stylistic notes, and the theory as a whole is outlined on the authorship overview.

Questions and answers

Who was Junius, and how was he identified?

Junius was the pseudonym of the author of letters published in the London Public Advertiser between 1769 and 1772, attacking the government of George III. Their author has been debated ever since. Alvar Ellegård's statistical study of 1962 supported the long-standing candidate Sir Philip Francis, a civil servant at the War Office, and his identification is widely accepted.

What is stylometry?

Stylometry is the statistical study of writing style, used to attribute anonymous or disputed texts. It measures features such as the frequency of common function words, sentence length and vocabulary richness. Modern methods work best on samples of several thousand words. In Shakespeare studies stylometry is used mainly to identify collaborators in plays such as Henry VIII and Pericles.

What survives in the Earl of Derby's own hand?

A number of letters signed "Will Derby" survive, some reproduced by A. W. Titherley in 1952, along with estate and legal papers. Titherley and later Derby supporters also attributed to him a manuscript copy of a poem printed in The Passionate Pilgrim. No play manuscript in his hand is known, although a 1599 letter reports that he was writing comedies.

Why did Nordling compare Derby with Edmund Spenser?

Spenser was a major poet of the same generation with a large surviving body of verse and a concordance to his vocabulary, so he made a convenient control. Nordling showed that many words and phrases in Derby's letters are absent from Spenser but present in Shakespeare. Critics would note that Spenser wrote archaizing verse, not letters or plays, which makes him an imperfect comparison.

Continue the voyage

A large Tudor manor house of brick and timber with lit windows, reflected in a moat among old trees under a crescent moon
Shakespeare and the authorship question

The real Shakespeare?

Read
London around 1600 across the Thames at night: houses on the old bridge with lit windows, church spires and small boats with lanterns
Shakespeare and the authorship question

The Establishment is unyielding

Read
A Renaissance castle with green copper spires on a headland above a narrow strait at night, a ship with a lantern passing below
Shakespeare and the authorship question

Stylistics, genealogy, etc.

Read