Document Type

Honors Thesis

Major

Mathematics

Advisor(s)

Jane Wang

Committee Members

David Hiebeler, Benjamin Johnson, Timothy Reagan

Graduation Year

May 2025

Publication Date

8-2025

Abstract

Ancient texts have long presented a problem in authorship accountability. Many have used statistics to approach authorship problems — from the Pauline epistles to the Federalist Papers — and this practice has come to be called stylometry. Our interest is the investigation of pseudepigrapha: texts that are falsely attributed to an author, usually a famous one. This problem is very common in ancient texts. Horsfall, a classicist, says that “in the ancient world, (...) you could measure fame by the accumulation of forgeries” [8]. Stylometry has enjoyed a rise in popularity in the digital age for combating disinformation and plagiarism, but its roots are in the study of ancient texts. The targets for our analysis are two pieces of potential pseudepigrapha from the poet Virgil. The primary target is a potential interpolation within Virgil’s Aeneid dubbed the Helen Episode. The episode is a chunk of 22 lines in the midst of the second book. The transmission of the Aeneid is generally sound, but the Helen Episode only appears several centuries after Virgil’s death [21] [7]. There are classicists on either side of the question of the authenticity of the episode. Our second target is a collection of juvenilia called the Appendix Vergiliana. These are poems attributed to a young Virgil, and are widely considered to be spurious [8]. We have investigated stylometric methods for authorship verification on ancient Latin poetry with Mahalanobis distance analysis and Support Vector Machines (SVMs), and applied them to the Helen Episode and the Appendix Vergiliana.

Download

Share