Digitally rebuilding a lost city
This story originally appeared in Yale Engineering magazine.
The ancient city of Dura-Europos, on the bank of the Euphrates River in present-day Syria, has long fascinated archaeologists and historians for its cultural diversity — Jewish, Christian, Mithraic, and other religious groups lived and worshiped close to each other. Even the graffiti that archaeologists uncovered revealed an impressive mix of languages — Greek, Latin, Parthian, Aramaic, Hebrew, and Middle Persian among them.
All of which makes it a rich area of study, especially since there are thousands of artifacts and documents from the site, which has been threatened by looting and conflict in recent years. The city, founded in 300 BC and abandoned in the third century AD, wasn’t considered a major metropolis, but historians say its artifacts could tell volumes about the everyday lives of people from that time and region. But because these artifacts are located in numerous locations and are often unmarked or have multiple labels, it is extremely difficult for researchers to study them.
Holly Rushmeier, the John C. Malone Professor of Computer Science, and Anne Chen, a postdoctoral associate at ARCHAIA, an interdisciplinary program at Yale for the study of ancient and premodern cultures, are working to change that. They recently received a $350,000 grant from the National Endowment for the Humanities to develop a digital archive of materials related to the archaeological site of Dura-Europos. They will create a virtual data cloud known as “linked open data” to bring together the disparate materials from this region. This will create a user-friendly interface that allows researchers to access the data, as well as add their own contributions.
“The linked open data will make it accessible, so you can search across all the collections and find the stuff that is really important,” Rushmeier said.
The digital archive will also make it easier for researchers to find what they’re looking for in the rather dense excavation records. For instance, they’ll be able to map out the thousands of objects in Yale’s possession and identify where on the site they were originally found.
“Then, from that, scholars can start looking at interpreting what that means about everyday life,” Rushmeier said.
And they hope to eventually create 3D reconstructions of the city based on the many images they have. For instance, she said, reconstructing spaces of worship like the Mithraeum of Dura-Europos, the Baptistry, and the synagogue could help researchers better understand how they were used.
Part of the challenge with the current collection can be chalked up to the convoluted process of excavating the site, which began 100 years ago. There were three phases of the historic excavations. First, a British team came in to do some initial exploratory work but fled within 24 hours because it was a war zone at the time. A Belgian archaeologist headed up a French expedition a couple years later, but it again became a difficult place to work. And five years later in 1922, Yale researchers teamed up with French archaeologists for a more in-depth, collaborative excavation.
The collection is particularly valuable to history because much of what we know about it isn’t through texts, but from the excavated artifacts. For instance, the director of the Yale excavation found evidence in the 1930s that Dura-Europos was the site of one of the earliest known instances of a gas attack. We only know this by what was found at the site, including a coin that revealed when it happened, and chemical tests done on excavated walls.
“They tunneled under the city, under the walls and set off gas that poisoned people,” Rushmeier said. “So, it’s an interesting place because a lot of what we know about it is from objects rather than written history.”
At the time, institutions responsible for an excavation would often split the materials among themselves, so about half of the material came back to Yale. Much of the rest is in Damascus, and the Louvre in France has a relatively substantial collection. Other artifacts, mostly gifts from Yale, are scattered at various North American institutions. Yale’s share of the collection is found at different campus locations. The Yale Art Gallery holds most of the objects and the historic photographs. Beinecke Library has about 100 of the objects. One cuneiform tablet that was found at Dura-Europos is in the Babylonian collection and can be found at the Peabody Museum.
“This is another reason why this is a great test case for linked open data, because even if we didn’t go beyond the Yale aegis, we can demonstrate how for the first time, we could make Beinecke materials speak to Peabody materials, speak to Yale Art Gallery materials,” Chen said.
Making things even more confounding for researchers is the turnover among curators over 100 years at all these locations, as well as all the different styles in record-keeping.
And then there are sites with multiple names. For instance, the first building that was excavated at the site is a house of worship that’s been known as the Temple of Bel, the Temple of the Palmyrene Gods, the Temple of Jupiter, and the Temple of the Oriental Gods. And that’s just in English. None of the materials that are in Western collections have ever been searchable in Arabic.
In three years, they hope to make the full collection searchable, with Arabic translations throughout. They also want to create a user-friendly website that allows visitors to easily toggle between English and Arabic, allowing researchers all over the world access to the collection to contribute their own expertise to the site.
They’re also aiming to use technologies developed in Rushmeier’s lab to build 3D models and other forms of geospatial reconstruction from photographs in the archive.
“So, if my group could do something to kind of presort the material in a way that then Holly and her students could batch the material and process it for 3D modeling, then great things could happen,” Chen said.
In addition to that project, Rushmeier and Chen also co-taught a course in spring of 2022, Introduction to the Digital Humanities for the Premodern World, which introduced students to various digital humanities methods and tools for studying the premodern world. Student teams created multidisciplinary projects related to the Dura-Europos site.
“We had a mix of computer science students and humanities graduate students taking the class together,” Rushmeier said. “We can do these interesting interdisciplinary courses that we’re really excited about in computer science, connecting things up to a lot of different areas.”
The class broke into teams, each having students from both computer science and the humanities. A humanities graduate student would pose questions or suggest topics to study, and then the computer science students developed techniques to move the project forward.”
One group developed a searchable database for the different examples of graffiti found at the site, with translation of the texts. Another group of students supplemented the content with information provided by the inscriptions for artifacts in museum exhibits. They also developed a QR code that visitors would scan with their phones and made use of the linked open data to access much more information about the items in the exhibit.
Strewn throughout the city of Dura-Europos were hordes of coins that residents hid whenever the city was attacked or some other crisis arose. A student team created a mapping device that shows their locations as well as some statistical analysis about the weights and origins of the coins.
Digitally organizing and reconstructing a collection from more than 2,000 years ago is an ambitious undertaking, but with the right technology and multiple disciplines working on it, Rushmeier is optimistic that it will soon be a valuable tool for researchers studying this remarkable site.
“With these multidisciplinary projects, we see how much we can push forward both to demonstrate things that can be done and surfacing these new questions that are opportunities for future research,” she said.
Media Contact
William Weir: william.weir@yale.edu, 203-432-0105