An amateur linguist and self-taught AI engineer has claimed to make progress in deciphering Linear A, the script of the Bronze Age Minoan civilization. The engineer used AI tools to test a hypothesis about one unknown word in a prayer inscription, which he linked to a Semitic root meaning 'to dwell' or 'to inhabit.' This led to the creation of a 408-word lexicon, suggesting Linear A is an extinct Semitic language. The AI-assisted approach allowed for rapid cross-checking of the hypothesis against thousands of characters, a task that would have taken a human months by hand. Source: arstechnica
AI excels at large-scale pattern testing, enabling it to verify hypotheses about linguistic structures across vast archives in minutes rather than years. It can also detect repeated sequences and predict missing characters in damaged inscriptions. Cross-lingual transfer, where a model trained on a known language infers patterns in an unknown one, is another strength. This method has been successfully applied to scripts like Ugaritic, an extinct Semitic language. However, statistical pattern matching alone cannot generate meaning without an anchor, such as a known language family or bilingual text. Linear A and Etruscan remain challenging because such anchors are missing or incomplete. Source: arstechnica
The AI model can learn which signs follow which and which words cluster together, but it does not understand the meaning of the symbols. This makes verifying AI-assisted claims difficult, as there are no native speakers or expert consensus to compare against. The small surviving corpus of Linear A, about 7,500 characters, further complicates validation, as any hypothesis can find scattered matches. AI remains a powerful tool for accelerating linguistic research, but it does not replace the need for a comparative anchor or rigorous human review. Source: arstechnica
