Researchers at Germany's University of Wurzburg have unveiled a groundbreaking artificial intelligence tool that is transforming how scholars read ancient clay tablets written in cuneiform, potentially unlocking countless forgotten secrets of the Hittite civilisation that once dominated Anatolia.
The tool, named Palaeographicum, functions as an extraordinarily sophisticated pattern-recognition system, automatically identifying individual variations in wedge-shaped characters pressed into clay over three millennia ago. Researchers say the development represents nothing short of a revolution in Ancient Near Eastern Studies.
"The Palaeographicum is radically changing our work," said Professor Daniel Schwemer, head of the Department of Ancient Near Eastern Studies at the University of Wurzburg. "It allows us to save thousands of hours."
The software searches through 70,000 digitised photographs containing over five million cuneiform characters, instantly locating identical or near-identical script formations. What once required three days of painstaking manual comparison across five clay fragments can now be completed in merely five minutes.
The Hittite civilisation, which flourished around 1600 BC in what is now Turkey, left behind a vast written record inscribed on clay tablets. However, most have fractured over the centuries, with their fragments scattered across museums worldwide. Before researchers could read complete documents, they faced the monumental challenge of correctly reassembling these broken pieces, a process that demanded identifying individual scribal characteristics in cuneiform's 375 distinct signs.
Cuneiform, unlike modern handwriting, presented peculiar challenges. Each of the approximately 400 scribes who wrote these tablets possessed recognisable stylistic quirks. Some pressed their styluses with such force that flourishes emerged, whilst others employed characteristic spacing patterns. Identifying these "fingerprints" became the key to piecing together fragmented texts.
The technology was developed through collaboration with the Technical University of Dortmund, building upon a five-year DFG-funded project called CuKa (Computer-assisted cuneiform analysis). The researchers are now continuously retraining the artificial intelligence, incorporating feedback from the global Hittitology research community.
Looking ahead, the team has set an even more ambitious objective: training the AI to automatically recognise individual scribes' handwriting styles. Should they succeed, scholars could compile a comprehensive "social history of Hittite writing culture," potentially revealing how individual scribes' work evolved throughout their professional careers and whether they wrote differently under pressure or at home.
The announcement has already garnered worldwide enthusiasm from the academic community, with Professor Gerfrid Muller noting that Hittite scholars "simply cannot do without" the portal housing this revolutionary tool.
