Scientists read 300-year-old letters without opening them

This is an Inside Science story.

Unopened letters more than 300 years old that were folded using mysterious techniques have now been read for the first time without opening them, a new study finds.

For centuries, before mass-produced envelopes started proliferating in the 1830s, most letters across the globe were sent using letterlocking, a method of folding letters to become their own envelopes. These intricate techniques also often served to help recipients detect if mail had been tampered with.

For example, during research in the Vatican Secret Archives, conservator Jana Dambrogio at the MIT Libraries unearthed Renaissance letters with odd slits and sliced-off corners. She discovered these were signs these documents were originally locked with a slice of paper slid through a slit and closed with a wax seal. Such letters could not be opened without ripping the paper, which would reveal to intended recipients if someone else had read the letters first.

After studying 250,000 historical letters, Dambrogio and her colleagues devised the first system to categorize letterlocking techniques, a kind of periodic table based on how these strategies creased sheets. "Letterlocking is a 10,000-year-old technology -- as long as people have tried to build security into their correspondences, you've had letterlocking, from cuneiform tablets all the way to bitcoin paper wallets," Dambrogio said.

Until now, scientists could only read these "letterpackets" by cutting them open, often damaging the documents. Although such work naturally focused on the letter's contents, this came at the cost of research into letterlocking itself, about which much remains unknown. "These folding sequences, they're just like a sneeze -- they're ephemeral," Dambrogio said. "The minute you open them, you lose evidence of them."

Now Dambrogio and her colleagues have devised a way to both read letterlocked texts without opening them and reconstruct the complex folds, tucks and slits used to secure them. "This is a pretty exciting major contribution to decades of work seeking to virtually unwrap artifacts," said computer scientist Brent Seales at the University of Kentucky, who did not take part in this research.

The scientists investigated the Brienne Collection, a postmaster's trunk holding more than 3,000 undelivered letters, including 577 letterpackets that were never opened. The letters were sent from all over Europe to The Hague between 1680 and 1706, the era that saw the Salem witch trials unfold, Newton reveal his laws of motion and gravitation, and Louis XIV move his court to Versailles.

The researchers first analyzed four letterpackets with high-resolution X-ray scans to generate 3D models of the documents. Next, they used a new computational algorithm to identify and separate different layers of the folded letters and recognize text written on them. Finally, the algorithm virtually unfolded the letters, not only making the writing visible, but also recording the crease patterns so the researchers could re-create the letterlocking process step by step.

This new strategy helped the scientists read unopened letterlocked text for the first time. For example, one unopened letter is a request from Jacques Sennacques, dated July 31, 1697, to his cousin Pierre Le Pers, a French merchant in The Hague, for a certified copy of a death certificate of a relative, Daniel Le Pers, perhaps due to a question of inheritance. The scientists detailed their findings online March 2 in the journal Nature Communications.

"It is unlikely that the letters in that trunk contain earth-shattering new revelations on politics, diplomacy, art or science," said historian Howard Hotson at St Anne's College in Oxford, England, who did not participate in this study. Instead, "what they promise to tell us is precisely the opposite -- how ordinary post circulated."

This new technique should also work for other collections of undelivered letters around the world, in addition to scrolls and anything else that unfolds flat. For instance, "there are a lot of old origami art pieces for which it was never recorded how they were made," said study co-author Erik Demaine, a computer scientist at MIT. "The idea of scanning them to be able to reconstruct how they were folded is really exciting."

Future research into letterlocking could potentially shed light on global patterns of culture and technological exchange, "as sophisticated letterlocking techniques are passed from one country, sphere of activity, or continent to another during the long period in which these techniques were in use," Hotson noted.

The scientists are making their techniques publicly available and open source for others to use and potentially improve. "We see this as the starting point for a lot of future research directions," said study co-lead author Amanda Ghassaei, an algorithm engineer at Adobe Research in San Francisco.

Inside Science is an editorially independent nonprofit print, electronic and video journalism news service owned and operated by the American Institute of Physics.