A Phenetic Approach to Selected Variants of Arabic and Aramaic Scripts

Abstract

This paper aims to introduce the phenetic method for processing paleographical datasets and evaluating their similarity relationships. The presented numerical taxonomic method was applied for selected varieties of the Arabic and Aramaic scripts. The phenetic model was evaluated by hierarchical clustering and—after applying multidimensional scaling—a centroid-based clustering method. The hierarchical clustering results were presented as dendrograms (phenograms), while the centroid-based results were given in 2- and 3-dimensional Cartesian coordinate systems. The obtained results demonstrate that the numerical taxonomy's phenetic approach is useful in describing the distances between different writing systems. The long-term goal of the research is to apply machine learning tools to clarify the relationships between the large number of Aramaic and Arabic script variants. This study belongs to the field of pattern evolution, in which machine learning methods of biological evolution are used to model evolving patterns (such as writing systems) over time.