A layered model for glyph identity and transformation in scripts

Abstract This paper presents a multilayer model for analyzing the identity and transformation of symbols within writing systems (hereafter referred to as scripts). The model comprises five interconnected layers: topology, visual identity, phonetic, semantic, and style. The topology layer defines the geometric and structural attributes of the glyphs of each symbol. The visual identity layer captures canonical features shared across glyph variants of a symbol. The phonetic layer links symbols to sound values, where applicable. The semantic layer situates symbols within their linguistic and cultural contexts. The style layer accounts for graphical variations introduced by instruments, scribal practices, and aesthetic conventions. Together, these layers constitute a general symbol model that can be applied across diverse scripts. As demonstrated through selected case studies, the model supports computational paleography, cross-script comparison, and the analysis of undeciphered inscriptions, advancing the formal modeling of script evolution and facilitating computational comparison and analysis of manuscripts.