Deciphering Undersegmented Ancient Scripts Using Phonetic Prior

Most undeciphered lost languages exhibit two characteristics that pose significant decipherment challenges: (1) the scripts are not fully segmented into words; (2) closest known language is determined. We propose a model handles both of these challenges by building on rich linguistic constraints reflecting consistent patterns in historical sound change. capture natural phonological geometry learning character embeddings based International Phonetic Alphabet (IPA). The resulting generative framework jointly models word segmentation and cognate alignment, informed constraints. evaluate deciphered (Gothic, Ugaritic) an one (Iberian). experiments show incorporating phonetic leads to clear gains. Additionally, we measure for closeness which correctly identifies related Gothic Ugaritic. For Iberian, method does strong evidence supporting Basque as language, concurring with favored position current scholarship. 1