How Codebreakers Decrypted a Trove of Long-Lost Letters Written by Mary, Queen of Scots

Though Mary Stuart and Elizabeth Tudor never met in person, the cousins—Mary’s grandmother was the older sister of Elizabeth’s father, Henry VIII—were natural foils to each other from afar. 

Elizabeth had grown up modestly, mostly away from court and sheltered from the long shadow of her disgraced (and deceased) mother, Anne Boleyn. Elizabeth was also believed to be a Protestant, and her Catholic half-sister, Queen Mary I, spent her own reign trying to knock Elizabeth down the line of succession, convert her, or at the very least make her marry a Catholic man. None of these attempts worked, and upon Mary I’s death in 1558, 25-year-old Elizabeth became England’s queen regnant.

Mary Stuart, by contrast, had become Queen of Scots mere days after her birth in 1542 and enjoyed a lavish childhood in France. Her sovereignty remained uncontested until 1567, when she was (wrongly) linked to the murder of her second husband and deposed in favor of her one-year-old son (the future James VI and I). Mary sought refuge in England, which was sort of a problem for the unmarried Elizabeth I—because Mary had a claim to the English crown, too. It was shaky at best: Henry VIII had specifically disinherited the Stuart line in his 1546 will, noting that succession should pass through his younger sister Mary’s heirs (the Suffolk line) if Elizabeth were to die without issue.

Despite this, many Europeans still considered Mary’s claim superior to that of her Suffolk cousins and felt she should’ve been acknowledged as Elizabeth I’s heir-apparent. Not to mention that Mary was Catholic, and therefore favored by the country’s Catholic contingent.

So, in the spirit of keeping your friends close, your enemies closer, and any throne-threatening cousins in the custody of a trusted nobleman, Elizabeth I tasked George Talbot, Earl of Shrewsbury, with guarding Mary Stuart at Tutbury Castle in February 1569. 

The ill-fated ex-queen remained under house arrest at one English estate or another until 1587, when she was beheaded for treason after being implicated in a plot to assassinate Elizabeth I. The damning evidence had come from coded correspondence intercepted and decrypted by Elizabeth’s secretary (and spymaster), Sir Francis Walsingham, and his cohort. Those weren’t the only letters that Mary sent during her lengthy stint in captivity: She wrote prolifically, and many of her missives are preserved in archives. 

Scholars have long assumed that plenty of Mary’s secret messages have been lost to history—but as it turns out, nearly five dozen of them were hiding in plain sight.

Climbing Up That Hill

Israeli computer scientist George Lasry, German musician Norbert Biermann, and Japanese astrophysicist Satoshi Tomokiyo are all avid cryptologists who habitually mine historical archives in search of encoded documents to decipher and study. While going through digital files from the Bibliotheque nationale de France, they unearthed 57 such documents catalogued among a number of early- to mid-16th-century Italian letters.

The trio rightly guessed that they were dealing with a homophonic cipher, in which more than one symbol represents certain common letters of the alphabet. Homophonic ciphers obstruct attempts at frequency analysis: It’s tougher to pinpoint the e’s, for example, if not every e is rendered as the same symbol.

To crack this one, the team employed a process known as hillclimbing. Basically, a computer randomly assigns the symbols to letters of the alphabet, decrypts the whole message, scores that decryption based on its intelligibility, slightly alters the cipher, and repeats the operation. The algorithm only keeps cipher changes that result in a higher score.

Things got a little easier once they realized that the messages were actually in French, not Italian. Words like temps (“time”) and prochaine (“next” or “upcoming”) revealed themselves, as well as promising fragments like catholi– and persecut–, all of which helped refine the cipher key. But still, computers could only help them climb so far up the hill, because the encoder had used more than just homophones to thwart potential codebreakers. 

For one thing, the cipher contained a number of atypical diacritics. “The hard part was understanding that the comma sign and the / sign were diacritics—to be read together with the preceding symbol, and changing its meaning, and not homophones representing specific letters,” Lasry tells Mental Floss. A plain c, for instance, represented the letter e. But c/ meant duc (“duke”), and c, meant homme (“man”).

There was also a puzzling symbol resembling a 4 with a tail, which they eventually worked out as a directive to repeat whatever symbol came before it. A curved exclamation mark, meanwhile, meant to delete the preceding symbol. In addition to those and other systemic quirks, the cipher boasted an extensive nomenclature: special symbols that stood for everything from people’s names to regular words (as illustrated by duc and homme above) and even parts of words. 

Unlocking the cipher’s many secrets happened through shared Google Docs and with lots of good, old-fashioned teamwork. “Since we come from different backgrounds, we all have our own way of looking at problems and solving them, so we could contribute ideas to each other,” Lasry says. “Norbert and Satoshi brought up angles and questions I would never have thought of, and that was the case for the others. Without this kind of collaboration, the results would have been incomplete and less reliable.”

Mary, the Mastermind

Lasry estimates that only about 30 percent of the text was legible when certain key details pointed them toward Mary, Queen of Scots as the possible writer—specifically, that it was an imprisoned woman who had a son—a theory that seemed all but certain when they came across the name Walsingham. To confirm it, they trawled through British archives for letters that matched theirs.

They managed to find decrypted versions of seven of them, which, aside from verifying Mary’s identity, helped them fill some holes in their own cipher key—and they could now use context clues from Mary’s history to fill in even more. The unknown symbol beside “my brother-in-law,” for example, could only be one of three men, and they landed on Francis, Duke of Anjou, as the right answer. The codebreakers even found crossover between their cipher and other known ciphers that Mary used to communicate.

It’s generally believed that this is the first time the other 50 letters have seen the light of day—and scholars are understandably jazzed to study them. Lasry, Biermann, Tomokiyo recently published their findings in the journal Cryptologia.

“It’s a stunning piece of research, and these discoveries will be a literary and historical sensation,” Tudor historian John Guy said in a statement. “They mark the most important new find on Mary Stuart, Queen of Scots, for 100 years.”

The letters, dated between 1578 and 1584, were nearly all written to Michel de Castelnau, England’s French ambassador and Mary’s conduit to her allies and other important contacts. Though her captors did usually allow her to send mail, Mary was well aware that it probably wouldn’t reach its destination unread.

“Please kindly apologize on my behalf to the Queen Mother [Catherine de’ Medici] that I don’t write to her directly, as I do not dare to commit anything not in cipher via this channel, and via the ordinary one, my letter would not fail to be discovered,” she wrote to Castelnau in 1583.

Keeping that back channel open and secret required ceaseless effort, which Mary frequently discussed in her letters to Castelnau. The messages are also a fascinating window into Mary’s response to—and role in—landmark political moments of the era, such as her panic-stricken appeals to France to recover her son after he was kidnapped by Scottish Protestants in 1582. 

But for all this discovery reveals about Mary’s years in confinement, it leaves one tantalizing question unanswered: How did her clandestine correspondence land in a file of older Italian documents in a French archive? It’s “definitely a direction for further research,” Lasry says. “And an important one, because if we can understand why the letters ended up there, maybe we may use that clue to locate other collections containing additional letters with the same cipher.”