Reading Ancient Scrolls with Modern Technology

Download the PDF

[classical music introduction]

James Cuno: Hello, I’m Jim Cuno, president of the J. Paul Getty Trust. Welcome to Art + Ideas, a podcast in which I speak to artists, conservators, authors, and scholars about their work.

Brent Seales: The idea is that you put the scroll in the machine and it does a pirouette. And as it turns around, the x-rays see what’s inside the scroll from every possible angle, 360 degrees, all the way around. And we can invert that and recover a complete representation of what’s inside, in three dimensions.

Cuno: In this episode, I speak with Getty curator Ken Lapatin and scientist Brent Seales about ancient Roman scrolls and the new technologies being used to read them.

[classical music fades out]

Mount Vesuvius erupted in the autumn of 79 AD, destroying the Roman cities of Pompeii and Herculaneum. A large and distinguished villa, likely owned by relations of Julius Caesar, was discovered buried at Herculaneum in 1750. Among the objects found within the villa were more than 1,000 carbonized papyrus scrolls, which gave the Villa its modern name, the Villa dei Papiri. Since their discovery, the scrolls have attracted a great deal of attention as it was hoped they could be read to reveal important unknown or lost ancient texts. Unfortunately, unrolling the carbonized scrolls without causing damage to them was next to impossible.

However, with the advent of the digital age, it has become possible to image the scrolls, introducing new approaches to reading them that could never have been envisioned when they were unearthed in the mid-18th century. To learn more about these recent developments, I spoke with Ken Lapatin, curator of antiquities at the Getty Museum, and Brent Seales, department chair and Alumni Professor of Computer Science at the University of Kentucky. Brent is also the director of the Digital Restoration Initiative and the inventor of virtual unwrapping.

So thank you, Ken and Brent, for joining me on this podcast episode. Some of our listeners may remember that two years ago, the Getty Museum organized and presented an exhibition on the Villa dei Papiri at Herculaneum, in Italy. Since then, the two of you have continued your work on the carbonized papyrus scrolls discovered there, applying new technologies to the investigation of their structure and content. And that’s what we want to talk about today, about what you’ve learned since the exhibition and what you’ve learned about the scrolls. So Ken, let’s begin with you. Contextualize the scrolls for us. Tell us what happened at Herculaneum in AD 79.

Ken Lapatin: In AD 79, unexpectedly to the ancient Romans who lived all around the Bay of Naples, Mount Vesuvius erupted. And it was a gigantic explosion of the mountain, which rained ash on Pompeii and poured volcanic mud down the side of the mountain, destroying the cities there. Herculaneum, located more closely to the volcano than Pompeii, was also destroyed and buried. But the volcano affected the site differently.

It was buried much more deeply under volcanic mud and debris; and thus, the preservation is much better at Herculaneum. And foodstuffs and natural products like papyrus scrolls were carbonized—flash fried, as it were—and thus, although damaged, preserved. And that’s what makes Herculaneum so very special.

Cuno: Where in Herculaneum were they found, and by whom were they found?

Lapatin: So just adjacent to Herculaneum, which was on the western coast of Italy, was a large luxury villa that has been named after the papyrus scrolls that were found there. We call it the Villa dei Papiri. And it was discovered by well diggers in 1750. And by tunneling deep underground, they first discovered marble floors, and then started extracting bronze and marble statues, digging deeper and deeper.

And after two years of digging these tunnels and extracting artifacts that went into the royal collections of the King of Naples, they found ancient scrolls that were rolled up. And at first, thy didn’t even recognize them, they were so blackened and charred. They were compared to a billy goat’s horn or other products of nature I won’t even describe; but you can imagine.

Which is that most of the scrolls are now in Naples, in the National Library. The vast majority of them. But a few of them are in Oxford and in Paris, because they were gifted by the kings of Naples to Napoleon Bonaparte and to the chaplain of the prince regent of England, who became George IV. This is how important the scrolls were at the time that they were found, that they were given as diplomatic gifts.

Cuno: Were there scrolls found elsewhere on the site of Herculaneum?

Lapatin: No. No. No scrolls, Jim. But we have found wooden tablets that were like the notebooks of antiquity, that have recessed panels that were filled with wax. And using a stylus, ancient writers could make notations and accounts and contracts.

And these wax tablets have been found at Herculaneum and at Pompeii. And the wax mostly is gone; it’s melted away. But sometimes people writing pushed too hard through the wax, and they’ve scratched the surfaces of the wood underneath. And that can be recovered, sometimes with the naked eye, but also using technology. But those tend to be not literary texts like we have from Herculaneum, but things like notes, contracts, and lists.

Cuno: So you describe the scrolls as being carbonized. So they’re blackened. And how did one discover that there was anything inside them worth seeing? Is it when it broke open? When it broke open, what did it look like?

Lapatin: They were charred to various degrees, Jim. How many existed and how many are preserved is really debatable. We think about 11- or 1200 have been found, but they’ve been broken, they’ve been fragmented. The inventoried fragments now number about 1800, but there are more fragments, so we don’t know the exact number.

And some of them were really blackened. Some of them were better preserved, so they could be read. There’s a great variety. And some of them were accidentally destroyed because they weren’t recognized.

But the story goes, one of them was dropped was cracked open, and Greek script was recognized within in. Most of them have black ink on black-charred papyrus. But some of them have black ink on sort of brownish papyrus. So there’re different degrees of legibility among the scrolls that have been opened. So some of the ink is visible to the naked eye.

And this excited everybody greatly and they brought them in to the king. And this started a competition among the courtiers for who could decipher and unscroll these papiri. And shortly thereafter, they enlisted a priest from the Vatican, Father Antonio Piaggio, who came down and made the first successful or semi-successful attempts to actually open the scrolls and try to read them. Some of the best scrolls were sureably destroyed by early attempts to read them. And some of them have, in fact, been opened, although very badly damaged.

Cuno: Were they exclusively Greek, or were there were also Latin?

Lapatin: Mostly Greek, but not exclusively. There are some Latin papiri, as well. And some of the Latin poetry gives us our dates, our latest dates for the functioning of the library, in the late first century BC and even early 1st century AD. But most of the scrolls are earlier. They’re Greek scrolls with Epicurean philosophy. Epicurus was a philosopher who was active in Athens in the late 300s and early 200s BC. And so there’s this historical library. And most of the scrolls are by a follower of Epicurus named Philodemus.

They’re very important, also, for their ancient content. Those that have been opened have revealed texts that were entirely unknown, entirely lost. Not the great works of Sophocles and Euripides and Aeschylus, but works of Hellenistic philosophy, of literary criticism, of authors whose names we knew, but works we did not have. And so that’s what has already been opened; but there are hundreds of scrolls that were too badly burnt, too badly carbonized, that can’t be opened. And the prospect of digital unwrapping gives us the opportunity not only of seeing what these texts are, but for seeing them undamaged, without the rips and tears that have resulted from physical unwrapping.

Cuno: And what do they tell us about the owners of the villa, that they should have such a library as that?

Lapatin: Well, they tell us a lot. First, they tell us that the owners, whoever they were, were interested in Epicurean philosophy and the questions that have, I think, occupied philosophers and many of us for time immemorial. What is it to live a good life? How to live without pain, without fear, how to be happy. What is enough to make us happy? But the texts also have literary criticism in them and very recondite arguments among philosophers.

So they tell us that the owners were interested in philosophy and Epicureanism. And we know this was true of a lot of wealthy Romans of the upper class during this period of the first century BC. And because so many of the scrolls are associated with Philodemus, and Philodemus, we know, was supported by a man named Lucius Calpurnius Piso, who’s most famous to us as the father-in-law of Julius Caesar, most scholars associated the villa itself with Piso and his son of the same name, who was a high imperial official under Augustus and Tiberius.

So we’re dealing with Romans of the highest rank, who are interested not only in governance in military affairs, but also philosophy. And the villa of the papiri, which is the model for the Getty Villa, also contained one of the greatest collections of art from the ancient world, in terms of statues in marble and bronze, ivories, wall paintings.

Cuno: Now we wanna get to Brent and talk about the technology that’s employed in this discovery and the unveiling of the papiri. But there were earlier attempts at this in the 18th century and the 19th century. Tell us about those very early attempts?

Lapatin: Once they realized there was Greek writing inside these damaged scrolls, they were eager to get at lost works of ancient literature. And they hoped for the great works of literature—lost plays by Aeschylus and Sophocles, the lost books of Livy. But the outer layers of the scrolls were the most badly charred. And if you think of an onion wrapped up. So they’d slice them down the middle or try to peel out the outer layers, breaking them. The layers would stick and the scrolls would be badly damaged. Father Antonio Piaggio, they thought he’d be down for a few months. He spent 40 years in Naples. But in that time, he only managed to unscroll about 19 scrolls.

Cuno: There was a machine-like form that I recall from the exhibition, that someone had developed as a means to unroll the scrolls, is that correct?

Lapatin: That was Father Piaggio’s machine, first developed in the 1750s, and then refined over decades, and then used through the 19th century. And that’s kind of a bizarre traction device, where the edge of the scroll would be found, and using animal glue, silk threads were attached to the edges. And then using he weight of the scroll, it was stretched just centimeters a day. This is why it took years and years to unscroll these book rolls. But he was meticulously opening the scrolls on what’s—If it were a human placed on this, we’d call it a torture device.

And these give us some of the best scrolls we have from the past. But in doing so, he damaged them. The layers would stick, there’d be rips and tears. Others attempts were even worse. Dip the scrolls in mercury, hoping the mercury would seep in and open the scrolls. This destroyed them. Put them in vacuums. This destroyed them. They got more and more destroyed. And Brent can talk about some of the more recent attempts, both the unsuccessful attempts of the 20th century, and now especially, the non-invasive attempts that he’s been at the forefront of.

Cuno: Well, let’s do that then. Brent, tell us about your background and how you got interested in the scrolls.

Seales: I’m a computer scientist, so I came to the scrolls from the point of view of digitization. And I think the 1990s really was the decade of a transition from, you know, looking purely at chemistry, at physical process, and at photography, and moving more toward what you could do with digital methods, and even noninvasive methods.

Some of those researchers in the ’70s and ’80s were looking at a little bit more sophisticated chemical means to try to pull layers apart, usually in small fragments; and then some photographic process that would document so that the puzzle could be put back together. But that quickly gave way, in the late early ’90s, to sort of a moratorium on the physical unwrapping, and a kind of looking forward to the digital age, where technology might actually be able to completely eliminate any physical handling and move toward a noninvasive analysis. And that’s about where I entered the scene in the Herculaneum story.

Cuno: Well, the papiri were not your first introduction to the possibility of using technologies to digitize manuscripts. What were the earlier attempts that you engaged in?

Seales: The papiri really are the most challenging thing because the form factor is a scroll. The material’s rolled up on itself. My introduction was with codices, proper manuscripts that had bindings and pages that could be turned. And what we were doing then—this would be in the mid-’90s—is that we began photographing open pages of beautiful illuminated manuscripts.

And then using different kinds of light sources to make those manuscripts appear differently, sometimes improving the contrast with ultraviolet or infrared light. And that became my entrée with those relatively straightforward applications, where you have something; it’s open in front of you, and you can take a photograph of it. And yet those photographs provided an enhancement that really started to move the field forward in how we were treating manuscripts, and ultimately scrolls.

Cuno: Did you get introduced to the scrolls in Herculaneum, or did someone tell you about the scrolls and then you went and worked on them?

Seales: We moved from manuscripts to the idea that we might be able to work on books that were completely closed, scrolls that were completely closed. But I did that from a technical perspective, without knowing about Herculaneum. A classicist by the name of Richard Janko, from the University of Michigan, who is a Herculaneum scholar—one of the foremost, actually—introduced me to the collection. And it was at that point that I realized the magnificence of what was there, and also the fragility, and the fact that some of the intact scrolls presented a real mystery that was definitely worth my time.

Cuno: Now, tell us about something called elemental analysis, what is it, and what it promises for our learning more about the scrolls.

Seales: Yeah. Another thing that’s emerged in heritage science over the past 30, 40 years is our ability to understand the chemistry, with very fine precision, of the things that have gone into the production of something like a scroll. You know, the inks, the papyrus and where it came from, and all of the kinds of trace contaminants that might be part of an artifact that can give it sort of a unique fingerprint or a unique footprint from the ancient world. So the ability to do that chemical decomposition, or that understanding of what is chemically part of the artifact itself, has grown by leaps and bounds, in terms of sensitivity, over the past, say, 20 years.

Cuno: Well, what’s the imaging aspect of that?

Seales: What’s really interesting about the decomposition of a chemical analysis is that you can map it onto a sort of positional grid. So if you look across the surface of a manuscript, you can see what each place on that surface looks like. You can also create an image from what the chemistry is at every point. And that chemical view, viewed as an image, gives you an idea of the different elemental composition of something that might be in a manuscript. How does the ink vary chemically, as you work your way across the page? What were the pigments that were used, and are they different from the first page and the last page?

Cuno: Yeah, I wanna get to this question about virtually unwrapping the scrolls. Now, who took the first attempt at that? And where is the study of this kind of technology being most advanced?

Seales: I think we were the first to try virtual unwrapping on Herculaneum. We’d done a number of experiments in the lab that were very successful. We made a pitch to Oxford, Friends of the Herculaneum in Oxford, about mid-2000s, to get access to scrolls and give it a try with real Herculaneum material. And so in 2009, we scanned, with a microcomputed tomography machine, two scrolls from the Institut de France, originally from Herculaneum. And that would represent the first time a completely three-dimensional acquisition had been made of an intact scroll. And from there we began the process of trying to do the virtual unwrapping.

Cuno: How do you do the virtual unwrapping?

Seales: It’s a software pipeline that requires first an acquisition that allows you to see all the way through ever page or every wrap of the scroll.

Cuno: What do you mean by acquisition? Is that an imaging term?

Seales: Yeah. The ability to unwrap something requires first some representation of it that allows you to see all the way through, and that’s where the tomography gives us a digital representation all the way through every thickness of every wrap of a scroll. So acquiring a representation from tomography is the starting point. And then every stage of the unwrapping from there represents a digital algorithm that operates on that original data.

Cuno: So how do I know that what I’m seeing is not compromised by being overlaid with another part of the scroll? In other words, how do you virtually distinguish between one layer of scroll from another?

Seales: Yeah. So the problem becomes segmentation, which is to find every one of the layers uniquely, identify it and represent it. The problem is to build a representation, not unlike taking a digital photograph and finding in the photograph, the things that are of interest—a head here, a dog here, a person there, a car there. Likewise, if you have an image of a set of wraps of a scroll, the trick is to find every one of those layers, identify it and represent it properly, ’cause that’s where the writing’s gonna be.

Cuno: Well, tell us how important the scrolls are for our understanding of not just the library itself, or what we call the library at Villa dei Papiri, but about prospects for other work in this field.

Seales: From a technical point of view, the scrolls are incredibly important, as probably representing the hardest kind of virtual unwrapping possible, and also the—maybe the biggest potential for discovery. I mean, the Herculaneum collection is the only complete library ever discovered from antiquity. I believe there’s still unknown treasures to be found in the material that is not yet unwrapped. So a technical solution would allow for us to extract more information from what’s there. But it would also motivate the idea of possibly further excavating the site to find if the library has been fully uncovered or if maybe there are more scrolls awaiting discovery.

Cuno: What about comparisons with other kinds of scrolls, if they exist, that are made of different materials and that may have existed in different kinds of climatic conditions, like high deserts or something? If the problem with the scrolls at the Villa dei Papiri is that they’re folded onto themselves or rolled onto themselves, are there other examples of scrolls or of papyrus elsewhere in the world that might be in situations that are similar, that might be helpful to you as you explore this at the Villa dei Papiri?

Seales: There are other collections. I would say that to my knowledge, Herculaneum is the largest collection in one place. But certainly, for example, the Dead Sea Scroll collection and the parts of that collection that are damaged and have resisted conservation techniques. We’ve done some work with those fragments and have found that in the area of the Dead Sea, there are a number of sites that produce amounts of material where this method could be applied.

It turns out that virtual unwrapping has creating a following for a lotta different materials around the world. Most recent one is letter-locked material where letters were folded together and interlocked, as a way of sort of protecting them, and then sealed with wax, with adhesive. Different material; but because the technology is advanced by looking at this material, it definitely will feed back into the application to something like Herculaneum, as the field matures and virtual unwrapping as a technique becomes more versatile and widely used.

Cuno: When we talk about the papiri as being folded onto itself, rolled onto itself, is there some structure within it that is wooden or metallic, on which the papyrus scroll is rolled? Or is that not the case?

Seales: No, it is the case. We’ve found a real surprise, in fact, that sometimes the scrolls we’re looking at are rolled onto what used to be a wooden dowel. And likely, that was how the scroll was stored and used. And then other times, we’ve found that there is no wooden dowel. They were simply rolled up like a piece of newspaper. I think that probably, the form factor was normally to have that wooden dowel there, and that was how they would’ve been placed in the library.

Cuno: Are there other circumstances where technologies, new technologies could be employed or are being employed that would be helpful for you, as you’re looking at the Villa dei Papiri scrolls? I’m just thinking about medical technologies.

Seales: In fact, I think one of the big breakthroughs has been the application of techniques that were thought to be really useful only in the medical realm, applying those to heritage materials. Tomography is one of those things, where you’re probably familiar, at the doctor, with going to a machine that surrounds you. And then when you come out, the doctor can see everything inside, sometimes in three dimensions.

Cuno: Is that a CAT scan? Is that what you’re describing?

Seales: Yeah. CT scan would be short for computed tomography scan. And we use the term micro for very high resolution, better than you can do with a human, because you can increase the dosage of the x-ray, since you’re dealing with materials that are not alive.

That tomography system in medicine is what we’ve applied to virtual unwrapping for scrolls. The idea is that you put the scroll in the machine and it does a pirouette. And as it turns around, the x-rays see what’s inside the scroll from every possible angle, 360 degrees, all the way around. And we can invert that and recover a complete representation of what’s inside, in three dimensions.

And using that three-dimensional representation, just like from the medical community, gives us the ability to do an analysis inside the scroll that you wouldn’t be able to do otherwise. So I actually think it’s very significant, the advances from medicine being applied to heritage materials.

Cuno: Where else in this country, in the United States, is this work being done?

Seales: Virtual unwrapping is fairly new. And the scanning happens at, usually, places that are known to be synchrotrons. So you’re probably familiar with the Archimedes Project and the work they did at SLAC. I think SLAC is an acronym that stands for the Stanford Linear Accelerator. And it’s an x-ray source that allows you to tune the x-ray to get much more precision in the analysis of materials than you would get from a bench source that might be in the lab or at a doctor’s office. So it’s an example of one of the synchrotrons nationwide that you could use.

There’s another one in Chicago. And then some of the beamlines at our national labs, they host the analysis of heritage materials and there’s definitely tomography going on at those places. The ones we’ve used are in Europe. The one in Oxford, and also in Grenoble, which serves the European community.

Lapatin: But part of your project entails, post-COVID, if I’m not mistaken, of taking a smaller machine to Naples and actually doing the scans there, right?

Seales: It is possible to us a mobile machine, not a synchrotron, to do some of this analysis. And yes, we do plan on engineering a machine that could be placed onsite. And that’s how the technology is evolving. Not possible 10 years ago, but now possible through the work of engineering, to make these machines usable onsite.

Cuno: Well, gosh, it’s an incredibly fascinating and complicate thing that you’re embarked on. Everything seems to be new in this field. So the prospect of discovering, making big discoveries is probably fairly high.

Lapatin: One of the things that’s fascinating to me is how this site, the Villa dei Papiri, since its discovery in 1750 and the discovery of the papiri in 1752, has really been a kind of laboratory for the development of archaeological science. Cutting-edge technology has been brought to the site in general, and the papiri in particular, from the very beginning. From Piaggio’s revolutionary, although still damaging, attempts to open to scrolls to Karl Weber, the chief archaeologist and engineer in charge of the excavation, making the first ever archaeological plan that plotted the find spots of not only the papiri, but the statuary in the villa and its gardens.

That this site just keeps giving and giving and giving, and that Brent is now using 21st-century cutting-edge technology on finds from the mid-18th century that date back to the last centuries BC and first centuries BCE. So we’re really using and seeing technology developing across time. There are a few clumps of scrolls that remain joined together. And when we brought two—we brought three scrolls from Herculaneum to the Getty Villa for the exhibition, with great care, because they’re so fragile, and with the wonderful cooperation of our colleagues at the National Library in Naples, which cares for them.

And as part of the process, just to make the shipping containers, we took the scans, preliminary scans made by Brent and his team of the exteriors, to make form-fitting shipping containers. And from the scans, we made 3-D prints that were used to test the shipping containers. And of course, the 3-D prints, modern objects, we could handle very easily and not worry about damaging them. And we discovered that two of the scrolls with adjacent inventory numbers, that have always been presented at different orientations, actually nested into one another. And this was a small discovery.

They must’ve been shelved together in the library and preserved for thousands of years. But with the excavations, they were numbered together, but they got separated, and we didn’t know that they really belong together, nesting in one another. And that’s a rather different use of facsimiles, a 3-D physical facsimile of just the exterior form of the scroll.

Cuno: So you’ve been working on these scrolls now for some time, some years. What have you learned so far? What have you discovered that is on the scrolls themselves or in the process, or that we’ve learned about the method they employed to roll the scrolls onto themselves and then to unroll the scrolls from themselves?

Seales: We think that we’ve finally developed a technical approach that is gonna let us see the text in all its detail. This is a really challenging problem. The interior is very damaged. And the ink and the chemistry of the ink is not very different from the chemistry of the papyrus, meaning that the subtle difference has to be teased out. And that’s where artificial intelligence and machine learning is giving us an avenue toward being able to extract the actual texts. So right now, on open fragments, we can confirm that we can see the text from computed tomography. But we’re right on the verge of being able to do that for the wraps that are inside all the intact scrolls. That’s where we hope in the next year or two, we’ll have a breakthrough.

Cuno: Well, what’s next for the two of you on this project?

Seales: I have in our database, a very high-resolution scan of two scrolls from the Oxford Beamline, the synchrotron, at the Diamond Light Source outside of Oxford. And we feel like in that data, we’re gonna be able to elicit text, text that no one’s seen before. So what’s next for us is that we’re using techniques from machine learning and artificial intelligence to be able to enhance our ability to amplify that data and hopefully, read that text. We’ve been working on that for the last year; and when we’re able, we’re gonna go back and improve the scanning method to hopefully make a breakthrough.

Cuno: Ken, what about you? What’s next?

Lapatin: Well, I’m very excited to see what Brent is coming up with and has already come up with. As an archaeologist, I’m interested in the whole site of the Villa dei Papiri. And as a Getty Museum curator working out of the Getty Villa, I’m constantly trying to understand all aspects of the villa—its owners, its art collections, its literary collections, the philosophical impulses of the people who lived at worked there. So as we learn more and more about the site and try to integrate these various threads, I’m excited to get a more holistic view of the lives and thoughts and activities of the ancient Romans there on the Bay of Naples.

Cuno: Well, thank you both very much for speaking with me today. This is a fascinating subject, and as you made it very clear to us, a very important subject, too. So we thank you very much for talking to us on the podcast.

Seales: Thank you very much, Jim.

Lapatin: Thanks, Jim.

[classical music fades in]

Cuno: This episode was produced by Zoe Goldman, with audio production by Gideon Brower and mixing by Mike Dodge Weiskopf. Our theme music comes from the “The Dharma at Big Sur” composed by John Adams, for the opening of the Walt Disney Concert Hall in Los Angeles in 2003, and is licensed with permission from Hendon Music. Look for new episodes of Art + Ideas every other Wednesday, subscribe on Apple Podcasts, Spotify, and other podcast platforms. For photos, transcripts, and more resources, visit getty.edu/podcasts or if you have a question, or an idea for an upcoming episode, write to us at podcasts at getty.edu. Thanks for listening.

[classical music fades out]