AI Deciphers Carbonised Herculaneum Scroll

Artificial intelligence has played a pivotal role in deciphering a carbonised papyrus scroll from Herculaneum. This artefact, one of over 800 recovered from the Villa of the Papyri, was rendered fragile and unreadable by the A.D. 79 eruption of Mount Vesuvius. The Vesuvius Challenge, a global competition, spurred efforts to read these scrolls without physical damage.

Contestants used machine learning to analyse X-ray computed tomography scans of the rolled papyrus. This non-invasive method aimed to detect minute variations indicative of ink on the carbonised surface. Early attempts in previous challenges focused on identifying the ink itself. The current breakthrough involved algorithms trained to identify specific patterns that represent Greek letters. These patterns emerged from areas where the iron-based ink subtly altered the papyrus structure.

AI algorithms identified distinct Greek text within the scans. The first word publicly confirmed was "πορφυρας" (porphyras), meaning "purple dye" or "purple cloth." This discovery marked a critical step in demonstrating the technology's capability. Further analysis revealed more extensive passages, confirming the scroll's potential contents.

The deciphered text includes a philosophical discourse. It is believed to be penned by Philodemus, an Epicurean philosopher who resided in the Villa of the Papyri. His writings comprise a significant portion of the Herculaneum papyri collection. The newly legible sections discuss sensory perception, specifically the experience of taste. This new information contributes to understanding ancient Epicurean thought and its nuanced views on bodily sensation.

This project represents a significant application of AI in humanities research. The technique could be extended to other damaged historical documents. It offers a pathway to unlock ancient knowledge previously considered irretrievable. What further ancient texts might AI render legible?