Colin Swaelens, Philology Meets Machine Learning: Digital Annotation and Semantic Modelling of Byzantine Greek Texts

Abstract

Diplomatic transcriptions of medieval texts pose particular challenges for digital annotation: their fragmentary preservation, non-standard orthography, and heterogeneous linguistic features complicate processing and limit the reuse of existing tools. To address these issues, we developed a full annotation pipeline for unedited medieval Greek, including part-of-speech tagging, morphological analysis, and lemmatisation. A manually created gold standard of 10,000 tokens was used to train and benchmark transformer- based language models, resulting in DBBErt, a system optimised for historical Greek. This pipeline not only improves the accuracy of linguistic annotation in low-resource settings, but also provides a foundation for more complex research questions. As a case in point, a semantic textual similarity benchmark was designed to investigate intertextual connections across Byzantine book epigrams, demonstrating how annotated data can enable new perspectives on meaning and reuse. Our work highlights both the potential and the limitations of existing annotation strategies: while transformer models significantly enhance processing quality, traditional embeddings remain competitive in some contexts, and software constraints require careful workarounds. More broadly, we will show how digital annotation fosters collaboration across philology, linguistics, and computational modelling, creating added value for the study of medieval texts and manuscript culture.

Practical information

This round table presentation will be given at Byzantium beyond Byzantium. 25th International Congress of Byzantine Studies, which takes places in Vienna on 24-29 August 2026.

Date & time: Saturday the 29th of August 2026, 08:30

Location: University of Vienna (Universitätsring 1, 1010 Vienna)

Floris Bernard, The Materiality of Reading Poetry in Byzantium

Abstract

This paper reassesses how Byzantine poetry was read, arguing that an exclusive focus on oral performance obscures the significance of its material, written contexts. While orality was undeniably an important factor, numerous poems reveal that the physicality of texts and the act of visual reading were equally central to their reception. The paper discusses verse disputes in the margins of manuscripts, invective exchanges on scrolls, and the status of the poem as an object in other written genres. Manuscripts emerge as spaces for intellectual debate and literary production. Acrostics and stichometric games highlight the physicality of poems as objects. Byzantine poems circulated as pamphlets, gifts, pedagogical tools with visual aids, and even devotional items. Thus, poetic objects were given various cultural, intellectual, and spiritual meanings. In sum, the paper proposes to see Byzantine poetic culture as a spectrum of practices: hearing and reciting, but also writing, annotating, seeing, and handling. Each act of poetic communication consisted of an interplay between orality and textuality/materiality (the latter best captured by the German term Schriftlichkeit).

Practical information

This lecture will be given at Byzantium beyond Byzantium. 25th International Congress of Byzantine Studies, which takes places in Vienna on 24-29 August 2026.

Date & time: Tuesday the 25th of August 2026, 15:30

Location: University of Vienna (Universitätsring 1, 1010 Vienna)

Kristoffel Demoen, The Database of Byzantine Book Epigrams, 2009–2026– 20??. Assessment and Perspectives

Abstract

The DBBE (www.dbbe.ugent.be) was created and developed thanks to three rounds of substantial funding, the third of which ends in 2027. In this paper, we will (1) briefly look back at the basic (and evolving) principles, choices and objectives; (2) describe the current state of the database (the second online version); and (3) look ahead to the near and further future, pointing out some challenges and possible priorities for the remaining period of guaranteed funding. It is not our intention to provide a demonstration of the DBBE, nor to give an introduction to book epigrams / metric paratexts. Rather, we are targeting an audience of occasional or regular users of the database, who are to a greater or lesser extent familiar with the corpus and who can give us critical and constructive feedback and suggestions.

Practical information

This lecture will be given at Byzantium beyond Byzantium. 25th International Congress of Byzantine Studies, which takes places in Vienna on 24-29 August 2026.

Date & time: Thursday the 27th of August 2026, 17:30

Location: University of Vienna (Universitätsring 1, 1010 Vienna)

Eleonora Lauro, Byzantine Book Epigrams and the Italo-Greek Scribe: Self-Representation and Textual Engagement in the ms. Messina, Biblioteca Universitaria, S. Salv. 133

Abstract

This paper focuses on a group of book epigrams featured in the manuscript Messina, Biblioteca Universitaria, S. Salv. 133 [Diktyon 40794], produced in Southern Italy in 962/963. The paper aims to demonstrate that these poems offer unique insights into the persons and communities involved in the writing and use of the manuscript, as well as into the personal interaction of an Italo-Greek scribe-reader with Byzantine texts. In the 10th century, despite the political instability of the Byzantine presence in Southern Italy (mainly due to Arabic raids), the cultural ties between this peripheric region and Byzantium persisted, as exemplified by the local book production and consumption. This paper aims to contribute to our understanding of this connection through the analysis of metrical paratexts, also known as Byzantine book epigrams. These poems, often located at the thresholds of manuscripts, function as a bridge between the material creation of the codices and the intellectual involvement of its audience. Italo-Greek scribes, patrons, and readers used these poems as tools for self-representation and to express their relationship with the texts they engaged with. A particularly compelling example of this dual role of book epigrams occurs in the codex Messina, S. Salv. 133, which contains works by John Chrysostom. This Italo-Greek manuscript features four metrical paratexts, all located at different folios. Through these poems, the scribe leaves traces of his involvement in the material production of the manuscript, such as his name and the date of completion. Moreover, he expresses his personal connection to John Chrysostom, revealing his motivation for copying and reading this author. I will perform a textual and contextual examination of these poems in order to reconstruct the socio-cultural identity of the scribe and to illustrate how Byzantine texts inspired active and intimate forms of intellectual and spiritual interaction in the Italo-Greek world.

Practical information

This lecture will be given at Byzantium beyond Byzantium. 25th International Congress of Byzantine Studies, which takes places in Vienna on 24-29 August 2026.

Date & time: Thursday the 27th of August 2026, 17:30

Location: University of Vienna (Universitätsring 1, 1010 Vienna)

Paulien Lemay & Kyriaki Giannikou, Exploring Formulaic Tradition Across Disciplinary Boundaries: Byzantine Book Epigrams and Beyond

Abstract

Byzantine book epigrams – metrical paratexts found in the margins of manuscripts – form a highly interconnected corpus (Ricceri et al. 2023). Beyond the socio-historical connections apparent through metadata (e.g. provenance), these poems are linked through a network of formulaic expressions. Yet tracing these connections is complex: formulae are not simply repeated but creatively adapted. From lexical substitutions to the reproduction of patterns in more abstract forms, formulaicity in book epigrams has drawn the attention of linguists and Natural Language Processing researchers (Giannikou et al. 2024). In this paper, we explore how the formulaic material of Byzantine metrical texts extends beyond the manuscript tradition into other material contexts, such as stone inscriptions and frescos. To examine the formulaic tradition beyond – and in comparison to – manuscripts, we focus on the corpus curated by A. Rhoby (Rhoby 2009-2014). Our study combines philological analysis with computational linguistics. First, we use convolutional neural networks for textual pattern mapping and combine the results with FAISS, a system for textual comparison. We then undertake a linguistic analysis of the similarities and differences revealed, to assess the extent to which the formulaic material across media coalesces, and to evaluate whether this approach can be extended to the broader Byzantine epigrammatic tradition. Ultimately, our paper presents our methodological approach and the resulting insights, demonstrating how computational models, combined with philologically informed linguistic research, enable scalable comparative approaches into textual reuse and adaptation across the diverse writing media of Byzantine culture.

Practical information

This poster presentation will be given at Byzantium beyond Byzantium. 25th International Congress of Byzantine Studies, which takes places in Vienna on 24-29 August 2026.

Date & time: Thursday the 27th of August 2026, 10:15-11:45

Location: Arkadenhof, University of Vienna (Universitätsring 1, 1010 Vienna)

Eleonora Lauro, “Book Epigrams” Bizantini e Cultura Libraria Italogreca

Abstract

Questo progetto ha come oggetto di ricerca i cosiddetti “book epigrams” bizantini attestati nei manoscritti greci prodotti in Italia meridionale tra il X e il XIII secolo. Data la loro stretta relazione con la dimensione materiale e quella contenutistica dei codici che li tramandano, questi componimenti costituiscono una risorsa significativa per lo studio di pratiche e dinamiche della cultura libraria italogreca.

L’analisi testuale di un gruppo selezionato di epigrammi è integrata dall’esame paleografico e codicologico dei manoscritti in cui essi appaiono. Uno strumento fondamentale per la ricerca è il “Database of Byzantine Book Epigrams”, che raccoglie dati testuali e contestuali di paratesti metrici trasmessi da codici bizantini datati tra il IX e il XV secolo.
Il progetto mira a ricostruire l’identità socioculturale di copisti, committenti e lettori italogreci che si servivano dei “book epigrams” per documentare il loro coinvolgimento nella produzione e nella fruizione dei manoscritti. Inoltre, lo studio di questi paratesti si propone di identificare specificità regionali ed eventuali tracce di interazioni interculturali.

Practical information

This lecture will be given at the Terzo incontro di studio dei dottorandi e dei dottori di ricerca in studi bizantini, which takes places in Rome on th 27th of February 2026.

Date & time: Friday the 27th of February 2026, 11:45

Location: Sapienza Università di Roma

Crash Course in Greek Palaeography 2026

The ERC project Ancient Corrections (AnCor) organizes a two-day course in Greek palaeography within the Research School OIKOS in collaboration with the Leiden University Centre for Arts in Society, Leiden Papyrological Institute, Leiden University Library and the Greek department of Ghent University. The course is intended for MA, ResMA and doctoral students in the areas of Classics, Ancient History, Ancient Civilizations and Medieval studies with a good command of Greek. It offers a chronological introduction into Greek palaeography from the Hellenistic period until the end of the Middle Ages and is specifically aimed at acquiring practical skills for research involving literary and documentary papyri and/or manuscripts. This course gives the unique opportunity to practice reading on original papyri and manuscripts from the collection of the Leiden Papyrological Institute and the special collections of the Leiden University Library.

Programme

The course is set up as an intensive two-day seminar. Four lectures by specialists in the field give a chronological overview of the development of Greek handwriting, each followed by a practice session reading relevant extracts from papyri and manuscripts in smaller groups under the supervision of young researchers. After each practice session, there is the opportunity to see a selection of original papyri and manuscripts presented by scholars with firsthand experience working with these objects.

Tuesday, 26 May 2026

09:00-9:15 Introduction to the course

09:15-9:45 Introduction to papyrology – Dr. Serena Causo

09:45-11:00 Lecture on papyri of the Ptolemaic and Roman period – Dr. Joanne Stolk

11:00-11:30 Coffee break

11:30-12:30 Reading practice with papyri of the Ptolemaic and Roman period

12:30-13:00 Display of papyri in the Leiden Papyrological Institute

13:00-14:00 Lunch break

14:00-15:15 Lecture on papyri of the Byzantine period – Dr. Yasmine Amory

15:15-15:45 Coffee break

15:45-16:45 Reading practice with papyri of the Byzantine period

16:45-17:15 Display of papyri in the Leiden Papyrological Institute

19:00 Dinner

The practice and display sessions are taught by Dr. Yasmine AmoryDr. Serena CausoDr. Joanne StolkDespina Borcea MA, Irene Chioni MA and Elise Hoekstra BA.

 

Wednesday, 27 May 2026

09:00-9:30 Introduction to codicology – Dr. Grigory Vorobyev

09:30-10:45 Lecture on Byzantine book scripts: From the first codices to the eleventh century – Prof. dr. Floris Bernard

10:45-11:15 Coffee break

11:15-12:15 Reading practice with Byzantine book scripts

12:15-12:45 Display of Byzantine books in the Leiden University Library

12:45-13:30 Lunch break

13:30-14:45 Lecture on Byzantine book scripts: The Comnenian and Palaeologan periods – Prof. dr. Andrea Cuomo

14:45-15:15 Coffee break

15:15-16:15 Reading practice with Byzantine book scripts

16:15-16:45 Display of Byzantine books in the Leiden University Library

The practice and display sessions are taught by Prof. dr. Paolo ScattolinDr. Juan Bautista Juan-LópezDr. Divna Manolova, Dr. Francesca SamoriDr. Grigory Vorobyev, Kyriaki Giannikou MA and Eleonora Lauro MA.

Practical information

The study load is the equivalent of 2 ECTS (2×28 hours). Participants will be asked to read secondary literature in preparation for the seminar (distributed several weeks before the course). Extra material will be handed out during the course in order to continue to improve your reading skills afterwards.

There are no fees for participation in this course. Lunches on both days and dinner on the first day are provided free of charge. Travel costs and accommodation in Leiden are at your own expense.

The course will take place at the following venue: University Library (Witte Singel 27, 2311 BG Leiden).

Registration

Please register by sending an e-mail with a short motivation (max. 250 words, including your background, research interests and why you would like to follow this course) to j.v.stolk@hum.leidenuniv.nl. Priority is given to OIKOS MA and doctoral students and those who did not have the opportunity to follow course(s) on palaeography before. Registration closes by the final deadline of March 1st, 2026. Successful applicants will be notified soon afterwards.

Eleonora Lauro, Byzantine Book Epigrams as Tools of Self-Identification in Italo-Greek Manuscripts

Abstract

This paper aims to explore the function of Byzantine book epigrams as sources of insights into the mentality and identity of persons and communities that shaped Southern Italian book culture between the 10th and 13th centuries.
While Italo-Greek manuscript production was mainly connected to monastic settings, secular clergy and laymen were also involved, albeit less frequently, in the material realization of codices. A glimpse into the motivations and socio-cultural identity of these individuals is offered by Byzantine book epigrams, i.e. metrical paratexts located at the thresholds of manuscripts.

Within these poems, scribes occasionally left traces of their engagement with manuscripts by recording personal details such as names, offices and designations. These occurrences are especially noteworthy because, as recently noted by I. Hutter (2023), anonymity was the standard condition for Italo-Greek scribes, especially in monastic contexts. Therefore, explicit self-identification stands out as a deliberate and meaningful act, potentially reflecting individual aspirations, or an intimate form of intellectual connection between the scribes and the content of the manuscripts they copied.

This paper seeks to demonstrate how metrical paratexts can contribute to a deeper understanding of the profile of Italo-Greek scribes. Drawing on data and metadata stored in the “Database of Byzantine Book Epigrams” (DBBE), I will perform a textual and contextual examination of selected book epigrams featured in Southern Italian manuscripts that are rich in prosopographical information. The goal is to reconstruct the socio-cultural identity of scribes and to shed light on the personal motivations that shaped their self-representation within these poems.

Practical information

This lecture will be given at the “2nd International PROSOPON Workshop“, which takes place in Munich on 10-12 December 2025.

Date & time: Wednesday 10 December 2025, 18:00

Location: University of Munich

PhD defense – Colin Swaelens, Verse by Verse: Modelling Semantic Similarity in Byzantine Greek Poetry

The defense will take place on Friday, 19 September 2025.

Ahead of the defense, two distinguished scholars in Ancient Language Processing, Barbara McGillivray (King’s College London) and Thea Sommershield (Durham University), will each give a talk starting at 14:00 in Paviljoen Vandenhove.

The defense itself will take place at 16:00 in Paviljoen Vandenhove, followed by a reception at 18:00 in the faculty library.

You can register until September 9 via this link: https://forms.gle/QXSqf7zkhSaY8GL28.

Colin Swaelens, Lemmatisation & Morphological Analysis of Unedited Greek: Do Simple Tasks Need Complex Solutions?

Abstract

Fine-tuning transformer-based models for part-of-speech tagging of unedited Greek text has outperformed traditional systems. However, when applied to lemmatisation or morphological analysis, fine-tuning has not yet achieved competitive results. This paper explores various approaches to combine morphological features to both reduce label complexity and enhance multi-task training. Specifically, we group three nominal features into a single label, and combine the three most distinctive features of verbs into another unified label. These combined labels are used to fine-tune DBBERT, a BERT model pre-trained on both ancient and modern Greek. Additionally, we experiment with joint training — both among these labels and in combination with POS tagging — within a multi-task framework to improve performance by transferring parameters. To evaluate our models, we use a manually annotated gold standard from the Database of Byzantine Book Epigrams. Our results show a nearly 9 pp. improvement, demonstrating that multi-task learning is a promising approach for linguistic annotation in less standardised corpora.

Practical information

This poster presentation will be given at ‘The 63rd Annual Meeting of the Association for Computational Linguistics‘, which will take place in Vienna, Austria from July 27 to August 1st, 2025.

Date & time: Monday 28 July 2025, 18:00-19:30

Location: Austria Center Vienna (Bruno-Kreisky-Platz 1, Vienna, Austria)