Detecting Style in Ancient Latin

Background Stylometry is a field that uses statistical and computational techniques to study the style of authors. Stylometry is used to address questions authenticity, authorship, and chronologies, among other questions. Most famously, Mosteller and Wallace used statistical techniques to determine the disputed authorship of the Federalist papers.1 More recently in 2015, stylometric techniques have been used to argue for the authenticity of a hitherto unaccepted work by Shakespeare.2 Stylometry has also proven useful when applied to more ancient texts in Latin and Greek. Thanks to initiatives like the Perseus Digital Library3 and Thesaurus Linguae Graecae,4 a large number of ancient Latin and Greek texts are available in digital form. Yet despite these facts, Classicists have remained reticent to adopt these new techniques as readily as others. The purpose of this project is to further demonstrate the potential benefits of statistical and computation techniques when applied to ancient, Western texts. Specifically, I will be performing stylistic analysis of 8 ancient Latin authors, building a classifier to determine the author of a given text.