The Scientists Who Are Deciphering 'Unreadable' Ancient Texts

The Lazarus Project has a secret to reading invisible ink: capturing it with multispectral imaging.

Sometime around the year 1491, the German cartographer Henricus Martellus produced an influential map of the world, which was likely used by Christopher Columbus on his 1492 expedition to the Americas.

Naturally, Martellus made sure to note all the wildest rumors about the locations he had charted out. Text over south Asia claimed that the Panotii people of the region had ears so big they could curl up and sleep in them, Dumbo-style. A cartouche over the Indian Ocean warned of "a sea monster that is like the sun when it shines, whose form can hardly be described, except that its skin is soft and its body huge," which experts think is a description of orca whales. Japan was labelled with the tantalizing note "precious stones are found on these islands."

The Martellus Map. Image: Henricus Martellus/Yale Library Archives

But time was not kind to these fantastic messages, and the vast majority of them were muted over the centuries by material degeneration. The words either faded out with time, or had been obscured by damage. We would never know about them at all if it weren't for the emergence of new techniques in imaging science over the last two decades, which can decipher damaged, vandalized, or otherwise unreadable texts with startling precision and accuracy.

"There's text all over the map," Roger Easton Jr., one of the key players in this fledgling field, told me over the phone. "But it had all faded; you couldn't really read much of anything."

Easton is one of the only people in the world who knows how to resurrect these lost writings from obscurity, which means he frequently ventures to fascinating destinations in order to study rare manuscripts. As a professor at the Rochester Institute of Technology and one of the leaders of the Lazarus Project—an organization of specialists involved in deciphering historical documents with multispectral imaging—he has collaborated with scholars, scientists, and other specialists all over the world.

"That's the best part," Easton told me. "Meeting the people is the best part. They are so appreciative when you pull something out."

These interdisciplinary teams are able to unlock these difficult texts by capturing them over a wide range of wavelengths. In much the same way that an X-ray image of the sky yields a different perspective on the universe than an infrared image, pseudocolor pictures of text can extract words that have been invisible to the naked eye for centuries. It's like The Da Vinci Code, only with fancy cameras and better dialogue.

"[The Martellus Map] was among the more fun ones we have ever done," Easton told me. "In some methods, you can see the scribe lines but you can't see any text at all. In other methods, the text pops right out. I know we didn't get all of it, but I think we got quite a lot of it."

"This is sort of the ultimate treasure hunt."

A comparison of the "keystone" cartouche on the map, before and after processing. Image: Lazarus Project/EMEL/Megavision/RIT/Yale Library Archives

Indeed, these efforts go way beyond the Martellus Map. From deciphering the field diary of Victorian explorer David Livingstone, written in berry juice, to investigating Thomas Jefferson's early drafts of the Declaration of Independence—revealing that Jefferson erased the word "subjects" and wrote in "citizens"—imaging science is racking up incredible historical finds.

Perhaps the most famous example of the field's potential was when it was applied to the Archimedes Palimpsest, a gorgeous parchment codex with an extraordinary history.

The book was originally a copy of one of Archimedes works written in Byzantine Greek by an unidentified scribe who lived in the tenth century. It was either made in Jerusalem, or was transported there at some point over the subsequent centuries. In either event, it was in Jerusalem, in the year 1229, that a Christian priest scraped off the codex's ink and bathed the parchment, then repurposed the erased pages as a liturgical prayer book.

The palimpsest changed hands many times over the centuries and became its own little moldy ecosystem, which further damaged the writing. But eventually, in 1906, portions of the undertext were recognized as the work of Archimedes by classics scholar Johan Ludvig Heiberg. Having translated many other works of Archimedes, Heiberg knew the mathematician's handiwork when he saw it, but until recently, only snippets of the beleaguered original copy could be recovered.

That all changed with a landmark project that ran from 1999 to 2008, in which Easton and his fellow digital decoders were able to use multispectral imaging to extract an enormous amount of the text that had been erased. That nameless scribbler who carefully transcribed Archimedes's work over 1,000 years ago did not labor in vain after all.

Archimedes Palimpsest bifolio 120v-121r, under normal light compared to UV illumination processed using principal component analysis to reveal the undertext. Image: Kevin Bloechl and Roger L. Easton, Jr.

"This is sort of the sappy way that I put it," Easton said. "The scribe that wrote this stuff was sort of the imaging scientist of his day, so I'm trying to collaborate with this person over a 1,000 year timespan."

To that point, the palimpsest turned out to contain ideas by Archimedes that had not survived in any other documents, including an entirely new treatise called "The Method of Mechanical Theorems." This work showed that Archimedes was beginning to work on very modern concepts such as actual infinity as well as the groundwork for calculus. If it weren't for Easton's collaboration with that nameless Byzantine scribe, we never would have probed Archimedes's genius to this extent.

Since then, the field has accelerated even further. "When we built the system for the Archimedes, we had a six-megapixel camera that cost $7,000," Easton told me. "You can now walk into a campus bookstore and buy a better camera than that for $350."

But though the technology is rapidly maturing, there is still a bottleneck in the field when it comes to qualified professionals. The skillset is in high demand, but it also happens to be very difficult to automate and disseminate to a global audience. (Overcoming this hurdle is central to the mission of the Lazarus Project.) According to Easton, the process of figuring out which wavelengths will coax text of any given page is still largely trail and error. "You are never certain what method is going to work," he said.

As a result, experts in this field are swamped with projects. "We do have this unfortunate tendency to bite off more than we can chew," Easton admitted.

When I spoke with him in December, he was a few weeks shy of a trip to Chartres to image manuscripts damaged by bombing during World War II. He's also working on the "New Finds" volumes at Saint Catherine's Monastery in Egypt, which contain multitudes of works that were recently discovered in a hidden chamber of the monastery, cut off from the main building after a structural collapse. It doesn't get much more "Indiana Jones" than that. Maybe if there were snakes or holy grails in the chamber too, but that just seems greedy.

Saint Catherine's Monastery. Image: Berthold Werner

All the while, new requests for imaging projects continue to be added to the backlog, from priceless historical artifacts to sentimental family heirlooms. As more people enter the field and the techniques are increasingly democratized, we have every reason to expect mountains of inscrutable writings to be rendered legible at last. Imagine the possibilities.

"It is very gratifying to be able to read text that had been feared lost forever," Easton said.