Historians have observed that some letters in the Armenian, Georgian, and Caucasian Albanian alphabets closely resemble characters in Ethiopic, the ancient script of the Horn of Africa. Until now, most comparisons depended on scholars visually inspecting the letters; however, this method is difficult to verify or reproduce. Researchers at San Diego State University turned to artificial intelligence to avoid these concerns with the comparisons.
In the study, published in Digital Scholarship in the Humanities, researchers trained an AI model on over 28,000 images of Ethiopic characters. The model then measured the structural similarity between these characters and letters from three other alphabets that developed in the Caucasus region of Eurasia. Notably, the Armenian alphabet showed the closest resemblance.
Shapes Without Context
The team designed the model to focus solely on each character’s shape. The AI did not have access to historical records, religious texts, geographic information, or cultural context. It analyzed only the curves, angles, straight lines, and overall structure of each character. After training on Ethiopic, the model compared these patterns to those of the Armenian, Georgian, and Caucasian Albanian scripts, then calculated their similarity using mathematical distance metrics.
The analysis showed that Armenian letters were the most similar in structure to Ethiopic, while Caucasian Albanian and Georgian were less so. As a control, the team included the Latin alphabet, which showed much lower similarity. This suggests that the patterns found among the Caucasus scripts are unlikely to be random.
“Our aim was to move beyond visual impressions that are difficult to test or replicate,” said Sam Kassegne, a professor of mechanical engineering at SDSU and the study’s lead investigator. “By making our criteria explicit and mathematical, we introduced an objective computational approach that is easily reproducible. We believe that this reproducibility is the key contribution of our method.”
Computation Meets History
The findings become even more interesting when viewed in historical context. The Armenian alphabet was created around 405 CE, at a time when Ethiopic was already in use and spreading through the region. Historical records mention Ethiopian travelers passing through Jerusalem, Egypt, and Syria during this period. Mesrop Mashtots, who invented the Armenian alphabet, also traveled in the Middle East at this time.
The AI model was not given any of this historical background. Yet it still found Armenian to be the closest structural match to Ethiopic, aligning with the period when contact between these cultures was most likely.
“What makes the research significant is that computational geometry and historical scholarship converge on the same scripts and time period,” said Daniel Zemene, an SDSU graduate student and the study’s first author. “The model had no access to historical records, yet it learned purely from visual and structural data and identified Armenian as the closest structural match to Ethiopic within the very timeframe historians have long debated. That convergence between computation and history is powerful.”
Similarity Does Not Always Indicate Influence
The researchers point out that their findings have important limitations. Just because two scripts look similar does not mean one was copied from the other. Writing systems can develop similar features on their own, or both could be influenced by another similar source. Throughout history, Greek, Roman, Persian, and Arabic scripts have shaped each other in ways that weren’t always direct or deliberate.
It is still unclear whether these similarities reflect direct cultural exchange, a shared but undocumented source, or another factor. For the first time, that debate can be tested with quantitative data rather than solely on visual comparisons. The numbers now support what scholars had long suspected, while the approach allows other researchers to test and reproduce the findings.
Austin Burgess is a writer and researcher with a background in sales, marketing, and data analytics. He holds an MBA, a Bachelor of Science in Business Administration, and a data analytics certification. His work focuses on breaking scientific developments, with an emphasis on emerging biology, cognitive neuroscience, and archaeological discoveries.