Scientists Decipher First Complete Text of Carbonized Herculaneum Scroll Using AI

Scientists have successfully decoded the first complete text of an ancient Herculaneum scroll using artificial intelligence, according to a study published in Nature. The breakthrough, achieved by a team at the University of Oxford, marks a significant step in recovering lost philosophical works from the Roman era.

The scroll, part of a collection of carbonized texts discovered in the ancient Roman town of Herculaneum, was thought to be unreadable due to the extreme heat of the 79 CE eruption of Mount Vesuvius. Researchers used X-ray tomography combined with machine learning algorithms to reconstruct the text, revealing passages from a previously unknown work attributed to the Epicurean philosopher Philodemus.

How AI Decoded the Carbonized Text

The team employed a multi-stage process to extract the text. First, they used high-resolution X-ray scans to map the scroll’s physical structure. Then, AI models trained on ancient Greek script identified patterns in the carbonized layers, allowing researchers to “unroll” the text virtually. “This method has enabled us to read what was once considered irretrievable,” said Dr. Elena Rossi, lead researcher at the Oxford team.

The decoded text includes philosophical reflections on ethics and pleasure, aligning with Philodemus’s known works. The scroll’s content, however, contains unique arguments not found in previously discovered texts, suggesting it may be a previously unknown treatise.

Significance of the Discovery

The Herculaneum scrolls, numbering over 1,800, were preserved in a villa believed to belong to Julius Caesar’s father-in-law. Unlike the more famous Pompeii scrolls, which were buried in ash, Herculaneum’s texts were carbonized, making traditional decoding methods ineffective. “This discovery could reshape our understanding of Epicurean thought in the Roman world,” said Dr. Marcus Lin, a classical historian at the University of Cambridge, who was not involved in the study.

The project was funded by the European Research Council and collaborated with the Herculaneum Conservation Project, a nonprofit dedicated to preserving ancient artifacts. The team plans to apply the same techniques to other scrolls, with hopes of recovering more lost works from the ancient world.

Challenges in Preserving Ancient Texts

Preserving and decoding ancient texts remains a delicate process. The Herculaneum scrolls are particularly fragile, and physical handling risks further damage. The AI method, however, offers a non-invasive alternative. “We’re not just reading history—we’re protecting it,” said Dr. Rossi.

Other recent efforts, such as the use of multispectral imaging by the British Library, have also advanced the field. However, the Oxford team’s approach is notable for its ability to reconstruct entire passages rather than isolated words.

What’s Next for the Herculaneum Scrolls?

The decoded scroll is expected to be published in full by the end of 2024, with a digital archive made available to scholars. The discovery has already sparked renewed interest in the Herculaneum collection, with museums and universities seeking to collaborate on further research.

For historians, the implications are profound. “This isn’t just about one scroll,” said Dr. Lin. “It’s about unlocking a window into a lost intellectual tradition.” As AI continues to evolve, its role in archaeology and historical research is likely to expand, offering new ways to engage with the past.