Digital Classics Online

Nomina Omina: Ancient Greek and Latin Proper Names in the Age of Artificial Intelligence

2025-09-23T14:02:38+02:00

This paper is the introduction to the volume Nomina Omina. Detecting and Preserving Ancient Greek and Latin Proper Names in the Age of Artificial Intelligence, which collects the proceedings of a workshop held at Leipzig University in June 2024, thanks to the support of the German Research Foundation (Deutsche Forschungsgemeinschaft). This introduction outlines the goals and contributions of the workshop and the volume.

Naming Latin Texts in the Thesaurus linguae Latinae Index of Sources

2025-07-29T08:53:43+02:00

This paper discusses a specific category of names in the Thesaurus linguae Latinae: names of ancient authors and works used as source citations. These names cover nearly all known ancient Latin texts until about 600 CE, including ca. 1,035 authors or text groups and ca. 2,880 identifiable works. They were most recently printed as the Index librorum scriptorum inscriptionum ex quibus exempla afferuntur (1990) and, in partly updated form, appear online. Currently efforts are underway to transform this into an open-access database as part of a larger digital transformation of the project. The Index includes information about the dates of authors and works, authority control data, modern editions, the existence of ancient translations into or out of Greek, sample citations, and other meta-data. The paper discusses theoretical and practical issues involving these names and their future data structure.

Daidalos: NER for Literary Studies on Latin and Ancient Greek Texts

2025-07-29T08:29:13+02:00

Literary texts offer a wealth of unstructured data that can be harnessed for data-driven text analysis through Natural Language Processing (NLP). Named Entity Recognition and Classification (NER) is a crucial initial step in this process, enabling the automatic identification of entities such as persons, organizations, locations, and dates. However, NER faces significant challenges, particularly with historical texts in low-resource languages like Latin and Ancient Greek, due to limited annotated corpora and the dynamic nature of language. This paper explores the evolution of NER from simple extraction to semantics-aware entity disambiguation and linking, highlighting the importance of multi-layer annotation systems to enhance data quality and model accuracy. The interdisciplinary Daidalos project aims to bridge the gap between Digital Humanities and Classical Studies by providing an NLP infrastructure that supports various data-driven research methods, among others NER. One of the project’s case studies demonstrates the potential of NER in Classical literary studies; this is accompanied by proposals on other NER related literary research questions, e.g. on authorship attribution and stereotyping. Additionally, the paper offers some thoughts about teaching NER, presenting a framework to assess the required level of digital literacies when working on a specific research question. Finally, it discusses the implications of generative AI and Large Language Models (LLM) on NER and NLP in Classics, emphasizing the challenges for independent research posed by the high costs and limited transparency of LLMs.

Opera Graeca Adnotata: Building a 40M+ Token Multilayer Corpus for Ancient Greek

2025-07-29T08:41:30+02:00

In this article, the beta version 0.2.0 of Opera Graeca Adnotata (OGA), the largest open access multilayer corpus for Ancient Greek (AG), is presented. OGA consists of 1,999 literary works and 40M+ tokens sourced from the canonical-greekLit, First1KGreek, and PatristicTextArchive GitHub repositories, which together host AG texts ranging from approximately 900 BCE to 1400 CE. The texts have been enriched with nine annotation layers: (i) tokenization; (ii) sentence segmentation; (iii) lemmatization; (iv) morphology; (v) dependency structure; (vi) dependency function; (vii) IPA transcription; (viii) composition date; and (ix) CTS structure. The layers are described by highlighting the main technical and annotation-related issues encountered. The corpus is released in the standoff formats PAULA XML and its derivative LAULA XML and is queryable online through ANNIS.

Automatic Annotation of Nomina Sacra

2025-07-29T08:49:54+02:00

Nomina sacra are a specific kind of named entities appearing in biblical manuscripts. Due to the large amount of biblical manuscripts, many questions about nomina sacra could not be answered to the present time. In order to use the methods of Digital Humanities for research questions on nomina sacra, they need to be consistently and accurately annotated. We report on our recent efforts on combining Handwritten Text Recognition (HTR) with annotation for biblical manuscripts written in Greek majuscule script. We reflect on the lessons learned from this work, especially on the technical aspects such as the available NER algorithms for classical languages, the performance of machine-learning based tools in comparison to rule-based annotation algorithms. We also discuss the pro’s and con’s of the approach we chose in our work.

Greek and Latin Proper Names in Georgian Scholarship: Epigraphic, Lexicographic and Encyclopedic Traditions, Their Standardisation and Digitisation

2025-09-23T13:54:10+02:00

Greek and Latin proper names are embedded across many Georgian scholarly traditions, including epigraphic corpora, lexicographic works, and encyclopedic projects. Their rendering into Georgian has varied over time, shaped by Byzantine, Russian, and European influences, and by differing translational practices. As a result, multiple versions coexist, often creating inconsistencies in scholarship and pedagogy. This article examines three major resources: the Encyclopedia Caucasus Antiquus, the corpora of Greek inscriptions discovered in Georgia, and the Orthographic Dictionary of Greek and Roman Proper Names. It traces the historical stages of translating Greek and Latin names into Georgian and discusses current attempts of Standardisation. Particular attention is given to ongoing digitisation initiatives, including adopting the Cadmus platform for the Digital Caucasus Antiquus. The article argues that digitisation and standardisation of proper names is not simply a technical matter but a cultural and linguistic imperative. For languages with limited global digital presence such as Georgian, the creation of structured digital resources is essential for safeguarding scholarly traditions and ensuring visibility within international digital humanities.

Hypotheseis, a Database of Named Entities Surrounding Greek Rhetorical Exercises

2025-07-29T09:05:41+02:00

The rhetorical exercises written in Greek from the Hellenistic to the Byzantine age constitute a wide corpus of texts that in the last two centuries have been the subject of few studies in comparison with many other areas of ancient Greek literature. As a result, there is no comprehensive list of Greek progymnasmata and declamations nor an overall study of the topics they cover and of the Named Entities they mention (mythological characters, historical persons, places, events, etc.) has already been published. The relational database Hypotheseis, founded in March 2024, is a Wikibase instance that aims to fill this gap, providing an easy way to describe the Greek rhetorical exercises and the Named Entities to which they are connected as structured data, available in CC0 license; a SPARQL endpoint allows making sophisticated analyses on the collected data. This paper presents the data model of the database, the corpus of texts it intends to catalogue, the work done so far and what could be done in the future.

Annotating Named Entities in the Trilingual Inscription at Ka’ba-ye Zartošt (ŠKZ)

2025-07-29T09:10:26+02:00

This study examines proper names in the trilingual inscription of Shapur I at Ka’ba-ye Zartošt (ŠKZ) located in Naqsh-e Rustam, Fars province, Iran. We introduce a corpus of Greek, Middle Persian, and Parthian versions of the inscription aligned at both sentence and word levels, using the Ugarit translation alignment tool. Through manual extraction and categorization, nearly 400 Named Entities (i.e., proper names) were identified and classified as persons (PER), locations (LOC), or location derivatives (LOCderiv). The paper addresses methodological challenges encountered during the alignment of the text, as well as the extraction and classification of Named Entities, including ambiguities in determining proper names, variations in how some names have been recorded across different versions, and complexities in maintaining consistency in categorizing names across various languages. Additionally, we highlight the value of the aligned corpus as a lexicographical resource beyond Named Entity annotation. All datasets, including the aligned versions of the text and the extracted Named Entities, are openly accessible via GitHub and Zenodo to provide a foundation for further historical and computational research. Lastly, we explore the possibility of adding further annotation layers and linking the corpus to other datasets.

LGPN-Ling for the Preservation of Greek Personal Names in a Digital Environment

2025-07-29T08:44:54+02:00

The aim of this paper is to illustrate the preservation of Greek personal names in a digital environment, going beyond the concept of ‘name entity’ and taking into account the morphological, morphosyntactic and lexical entities involved in the analysis of these personal names. To achieve this, I will focus on the output of the LGPN-Ling project, directed by Sophie Minon: the LGPN-Ling search site and the printed volume of Lexonyme. I will first examine the structure of the website and its conception, taking into account the uncertainties that can arise in the analysis of personal names and that can be difficult to convey in a clear and concise manner. I will then use representative examples to illustrate how the structure of the LGPN-Ling site allows Greek personal names to be preserved in a digital environment not only as nomina omina, but also as even smaller entities such as morphemes and lexemes. Indeed, each entry in the database is connected and linked to the others in all aspects of its analysis. This includes gender, geolinguistics, bases, suffixes, morphosyntax, semantic and even lexical correspondences in Greek. Names are thus linked on the basis of one or more of these features, creating a complex network of interrelationships between the registered forms.

Altinum: a Wikidata Project for Digital Epigraphy and Prosopography

2025-07-29T08:57:57+02:00

Altinum is the first Latin epigraphy project to be hosted on Wikidata, the collaborative database maintained by the Wikimedia Foundation. Launched in 2024, the project involves importing information about epigraphic artefacts from the Roman period originating in Altinum, a municipality in eastern Veneto, being also the first attempt of a digital prosopographic corpus of the municipality. The data have been sourced from EDR, EDCS, analogue catalogues, and unpublished theses. Once the data are integrated into Wikidata, users can formulate queries to generate graphs and tables, obtain statistics, and reconstruct family trees. As Altinum demonstrates, this approach not only expands the epigraphic corpus and prosopographical data but also makes it increasingly accessible in a collaboratively editable, multilingual and interdisciplinary database with the possibility of highly customisable queries.

Digital Mapping of Toponyms in Paradoxographical Texts: The Case of the Paradoxographus Florentinus

2025-07-31T08:22:38+02:00

The article presents the first results of a digital project devoted to the study of ancient paradoxography, a literary tradition consisting of collections of strange-but-(supposedly-)true phenomena concerning the natural world and the human sphere. The article is structured in two parts. First, we describe the first steps towards the creation of a relational database of ancient paradoxography, in collaboration with Trismegistos (TM Paradoxography). Second, we focus on the possibilities and limitations of the digital study of the toponyms mentioned in ancient paradoxography, by discussing the digital annotation of the toponyms mentioned in the paradoxographical collection known as the Paradoxographus Florentinus. This case study shows that the digital mapping of toponyms can be a valuable research tool to better understand the structure of paradoxographical texts and, more generally, the geographical horizon of ancient paradoxography.

More than Names? Challenges and Opportunities for Ancient Named Entity Recognition

2025-07-29T09:01:34+02:00

This paper focuses on the conceptual challenges of Named Entity Recognition and Classification for Ancient Greek and Latin texts. It examines the shifting definitions of ‘name’ and ‘named entity’, their changes over time, the overlaps and differences between them, and shows how their use is often flawed by implicit assumptions on naming mechanisms in language and culture. It then offers examples of ancient place-naming practices that may challenge these assumptions, highlighting the limitations of current vocabularies and standards, and pointing to the need for a domain-specific approach to the problem of Named Entities in ancient languages.

The NIKAW Project: An Infrastructure of Texts, Entities and Language Models to Study the Circulation of Knowledge in the Ancient World

2025-07-29T08:12:20+02:00

This paper presents the foundational work of the interdisciplinary project NIKAW (Networks of Ideas and Knowledge in the Ancient World), which aims to analyse social networks in ancient Greek and Latin texts through mentions of historical figures. As a critical first step, we address the challenge of Named Entity Recognition (NER) for these languages by leveraging transformer-based models enriched with domain-specific knowledge. Our experiments highlight data sparsity and annotation inconsistencies as key bottlenecks for model performance. In the second phase, we introduce a pipeline for Named Entity Linking (NEL), utilizing the Wikisource edition of the Pauly-Wissowa Encyclopedy as a knowledge base. We detail the creation of silver-standard (automatically annotated) and gold-standard (human-verified) training datasets, and report preliminary results from fine-tuning the BLINK model for NEL.

Detecting Eastern and Western Names in the Latin Corpus of the SERICA Project – With Special Regard to the Confucius Sinarum Philosophus (1687) as a Case Study

2025-07-29T08:22:04+02:00

This paper aims to describe the role of ICT within the SERICA (Sino-European Religious Intersections in Central Asia. Interactive Texts and Intelligent Networks) Project, especially by focusing on the corpus of Latin texts we are progressively building. Particular attention will be paid to the annotation of Named Entities (NEs) through Recogito from a very peculiar 17th century Latin text, entitled Confucius Sinarum Philosophus (CSP) and edited by the Jesuits Prospero Intorcetta, Christian Herdtrich, François de Rougemont, and Philippe Couplet in 1687. Despite including the translation of three Confucian texts (Daxue, Zhongyong, Lunyu), the CSP contains various references to Graeco-Hellenistic and Roman literature, and this comes as unsurprising since the Jesuit Ratio Studiorum (1599) was indebted to pagan classical authors. Yet the reception of ancient Latin literature can be further investigated by resorting to digital technologies. The annotation and the extraction of NEs allow in fact to take into account an extensive amount of data and to establish a first mapping concerning the impact of classical antiquity on the CSP, so as to detect which authors were mentioned more often and to reflect on their pattern of distribution within the work.