While the main applications of resources and tools for sentiment analysis typically fall within the scope of fields like customer experience and social media monitoring, there is an increasing interest in extending their range to texts written in ancient and historical languages. Such interest mirrors the substantial growth of the area dedicated to building and using linguistic resources for these languages, which are essential for accessing and understanding the Classical tradition. In this talk, we will present the methodology we followed to create and evaluate a new set of Latin sentiment lexicons, and the process of inclusion of a prior polarity lexicon of Latin lemmas in a knowledge base of interoperable linguistic resources developed within the ERC project “LiLa: Linking Latin”. We will discuss the main challenges we face when working with ancient languages (e.g., lack of native speakers, limited amount of data, unusual textual genres for the sentiment analysis task, such as philosophical or documentary texts) and we will describe two use cases underscoring the importance of an interdisciplinary approach combining computational linguistics, semantic web and humanities practices.

Invited talk at the "Sentiment Analysis in Literary Studies" workshop organised by the Zentrum für Informationsmodellierung of the University of Graz (Austria).