Author(s): Martín-Arroyo, Daniel (CEIPAC-University of Barcelona) – Romanowska, Iza (Barcelona Supercomputing Center)
Imagine having 2000 inscriptions on roman amphora at your site. Each of them is unique but they do seem to follow some syntactic patterns (for example, “a name, followed by a date, followed by a number”). But how do you make any sense of them? Can these pat-terns be quantified? Do they group into possible ‘cargos’ or specific products? How are they related to the location of each amphora sherd within the city?
Here we present a Python-based software tool that compares epigraphic inscriptions to each other using the edit distance meas-ure, calculates clusters based of this similarity and visualises them as cluster maps and dendrograms. It can also look for correla-tions between the defined clusters and independent variables at different thresholds.
In short, this is a tool that aids the analysis of written sources by quantifying their syntactic similarity and providing visual and sta-tistical methods to aid the interpretation.
To showcase its functionality we applied the edit distance tool to a dataset from Pompeii comprising tituli picti found on amphoras throughout the city. The results show strong differences in the syntactic attributions between the latin and the greek worlds as well as a number of previously unknown spatial patterns.