Applications of Digital Analysis Techniques to Classical Languages and Literatures

Abstract

This article offers a survey of recent studies on the application of Natural Language Processing techniques to classical philology, with particular emphasis on Latin epic poetry of the 1st century AD. Although originally designed for the analysis of modern languages, the algorithms developed in computational linguistics over the past three decades – ranging from Latent Semantic Allocation and Latent Dirichlet Allocation to the more recent Transformer models based on Neural Networks – have transformed intertextual research through the practice of Topic Extraction. Particular attention is given to cases aimed at detecting situational correspondences even in the absence of co-occurring vocabulary. The examples discussed confirm both the efficacy of this approach and the convergence between ‘traditional’ philological methods and computational-linguistic techniques.