News, Events, CFPs, Jobs and Funding, Reading Groups

Online, 12 November 2026, 5-6.30 pm (UK time)

Nicholas De Sutter, KU Leuven: From Stage to Page to the Digital Age: Exploring AI as a Neo-Latinist

Andrea Peverelli, Syddansk Universitet: 'Words that Betray a Confession: Detecting Faith in Neo-Latin Drama, from Classifiers to Large Language Models'

Abstracts of the talks are below. If you would like to attend, please email Alison Shell (a.shell@ucl.ac.uk) for a link.

Nicholas de Sutter, 'From Stage to Page to the Digital Age: Exploring AI as a Neo-Latinist'

As a Neo-Latinist rather than a digital humanities specialist, I have approached AI primarily as a practical research tool, exploring how recent developments can support more traditional forms of scholarship. In this presentation, I reflect on my experiences using AI within my research on manuscript drama from the Southern Low Countries.

One area in which AI has already had a profound impact is the transcription of early modern texts. Advances in both Optical Character Recognition (OCR) and Handwritten Text Recognition (HTR) have greatly accelerated the production of working transcriptions of Neo-Latin printed books and manuscripts. Drawing on examples from my own research, I will discuss the combined use of specialised platforms such as Transkribus and more general AI tools in the process of transcribing and preparing early modern theatre texts for further study.

The presentation will also consider the next stage in the workflow: the transformation of digitised texts into structured, machine-readable research data. In the context of the ongoing development of NeoLatDraCor, a new Neo-Latin subcorpus within the DraCorLink opens in a new window ecosystem, I am experimenting with ways in which AI can assist in converting manuscript and printed playtexts into TEI-encoded texts suitable for digital publication and analysis.

Rather than offering a technical showcase, this talk provides a practitioner’s perspective on the opportunities, limitations, and challenges of integrating AI into the daily work of a humanities researcher. It asks what these emerging technologies can realistically contribute to the study of Neo-Latin literature and theatre, and where human expertise remains indispensable.

Andrea Peverelli, 'Words that Betray a Confession: Detecting Faith in Neo-Latin Drama, from Classifiers to Large Language Models'

Neo-Latin drama rarely addresses theology directly, and yet its language carries the fingerprint of its author’s faith. Two plays built on the same classical models, one written by a Catholic and one by a Protestant, can differ sharply in vocabulary, the choice of a word for love, judgement, or duty already carrying an underlying confessional weight. Working from a large corpus of early modern Catholic and Protestant drama, I show that a computational model can learn to tell the two confessions apart from word choice alone with great precision. The vocabulary that marks each side reveals that Catholic drama clusters around kingship, militancy, and sacred space (Caesar, rex, templum, Christus), while Protestant drama favours doubt, moral rectitude, and inward struggle (dubius, crimen, recte).

In the second part of the talk, I turn to a general-purpose AI system (a language model with no training specific to this material) and ask whether it reaches similar conclusions when faced with the same texts, and why. The comparison is a way into a broader question for Neo-Latin studies: what does a tool built and trained for one text and one question see that a general, all-purpose model does not, and vice versa?