Gábor Hosszú
Abstract Pattern systems are abstract communication frameworks composed of symbols, syntactic rules, and semantic structures. This study introduces the theoretical foundations of Deep Feature Phylogenetics as a model-free approach to the evolution of symbolic systems. Unlike traditional DNA-based phylogenetics, where substitution models are well-defined, the evolution of human-made patterns—such as writing systems—often lacks a predefined probabilistic model for state transitions. By situating scriptinformatics within the broader domain of Applied Computer Science, this paper proposes a framework to model structural evolution by representing pattern systems as graph sequences. We introduce the weighted holophyletic index (WHI) to quantify the clarity of these evolutionary lineages. The methodology is validated through the analysis of historical graph sequences (inscriptions), demonstrating that Deep Feature Phylogenetics can effectively reconstruct evolutionary dynamics even when the underlying transition probabilities are unknown.