Digitizing Antiquities: A 3D Photogrammetry Placement at the Institute of Classical Studies

Posted on 2026-07-06 by Gabriel Bodard

Digitizing Antiquities: A 3D Photogrammetry Placement at the Institute of Classical Studies

by Kiwa Chen and Kei Dai

During a two-week placement in the Senate House MakerSpace, University of London, in April and May 2026, my partner Kei and I focused on the 3D digitization and documentation of antiquities held by the Hellenic and Roman Library at the Institute of Classical Studies (ICS). Fulfilling a core component of our MA in Digital Humanities at UCL, this project allowed us to step out of the classroom and directly apply computational methods to the tangible preservation of archaeological heritage.

The ICS Library holds a fascinating collection of over 150 small antiquities, primarily Greek and Roman ceramics. These artefacts were bequeathed to the Institute by Dr. Victor Ehrenberg in 1976 with a clear and generous intention: they were not to be locked away, but rather used actively for handling and teaching. Today, this ongoing digitization initiative honors Ehrenberg’s vision. By producing high-quality 3D models of the collection using photogrammetry, we are able to upload these artefacts to the ICS Sketchfab account. This project not only meets the original mandate of teaching and handling but dramatically increases global access to, and engagement with, the Library’s collection.

Settling into the MakerSpace Workflow

Under the attentive and expert supervision of Dr Gabriel Bodard (Gabby), we were introduced to the foundational workflows of 3D modelling. Our digitization took place at the Senate House MakerSpace, a collaborative environment managed by the Digital Humanities Research Hub. We were equipped with a Nikon D850 digital SLR camera to capture high-resolution images, which we subsequently processed using Agisoft Metashape Pro.

Over the course of the two weeks, our principal tasks were threefold: photography and photogrammetric processing, meticulous documentation of our workflows, and archiving the resulting images and models. In total, we successfully modelled three distinct artefacts: two fragile ceramic sherds and one stone statuette.

Week One: The Ceramic Sherds and the Sponge Solution

Photography of a fragmentary ceramic potsherd, as described in the text.

Figure 1. Ehrenberg 18. Mycenaean Sherd

During our first week, Gabby thoughtfully suggested that we start with slightly simpler items to familiarize ourselves with the equipment, lighting, and software. Kei and I selected two ceramic fragments: the “18. Mycenaean sherd” (Figure 1) and the “16 Black-figure Sherd” (Figure 2).

This artefact is a light tan pottery sherd, possibly dating to the second Late Minoan period (L.M.II?, 1470-1420 BC), originally found at Tiryns in 1925. Measuring 68mm in height, 51mm in width, and 6mm in depth, the obverse is decorated with faded brown and red horizontal stripes. The reverse bears the fascinating handwritten inscription “Tirynes.”

Photograph of a black-figure potsherd, as described in the text.

Figure 2. Ehrenberg 16. Bf. Sherd

This black-figure sherd was acquired in 1925 as a gift from a shopkeeper in Athens. The obverse depicts an animal, likely a boar, with some paint loss across the body, while the reverse is glazed entirely in black. Interestingly, the item comprises two fragments visibly adhered together. Judging by its color, the adhesive appears to be animal glue, though the date of this historical repair remains uncertain.

See the 3D models in Sketchfab:

18. Mycenaean sherd by ICS Ehrenberg Collection on Sketchfab

16 Bf. sherd by ICS Ehrenberg Collection on Sketchfab

Challenges and Problem Solving

Our immediate challenge with these fragments was their physical shape. Unlike flat-bottomed objects, these sherds could not stand upright on the turntable. To capture a full 360º view, we needed a way to support them without obstructing the camera or damaging the 3,000-year-old terracotta.

Gabby helped us devise an ingenious, non-destructive solution: we carved a small groove into a piece of soft sponge. By gently resting the fragment within this groove, we could securely hold the item in place. We photographed the sherd in one orientation, then flipped it upside down in the sponge to photograph the reverse side. This allowed us to generate two separate “chunks” (one for the top, one for the bottom) in Agisoft Metashape. Figure 3 shows this setup in action.

Photograph of a ceramic fragment propped on a black sponge block and sitting on a white, electrical turntable in a photographic light box, as described in the text.

Figure 3. A ceramic sherd resting in the sponge support on the turntable during image capture

However, as we quickly discovered, relying on the software’s automatic chunk-merging feature rarely yields perfect alignment. To resolve this, we had to manually align the chunks point-by-point by adding corresponding markers across both halves. Yet, even with meticulous manual alignment, the two chunks rarely overlapped flawlessly, often leaving extraneous overlapping geometry where the halves met.

To refine the final piece, we had to use the selection tool to manually delete these redundant parts. This cleanup process required extreme spatial precision. We learned that finding the exact right angle to view and select the geometry was crucial; we had to constantly rotate the merged model in the 3D space to identify a “safe” perspective where we wouldn’t accidentally highlight the actual object. During our early attempts, failing to find this specific angle meant we occasionally deleted too much, resulting in missing gaps in the model itself. This iterative trial-and-error taught us that careful digital cleanup is just as critical as the initial photography.

Furthermore, the merging process faced additional hurdles with the 16 Bf. Sherd. The black glaze on the 16 Bf. Sherd proved highly reflective. In photogrammetry, inconsistent reflections across different photographs confuse the software’s alignment algorithms, as the “highlights” appear to move across the object’s surface. Consequently, our first attempts at merging the chunks resulted in slight blurriness and less-than-perfect geometry.

Week Two: Tackling the Stone Statuette

Photograph of a Romano-Egyptian ceramic baboon figurine, as described in the text.

Figure 4. Ehrenberg 30. Sitting Baboon

Armed with the hard-earned lessons regarding lighting and reflection from week one, we challenged ourselves in week two with a significantly more complex object: the Sitting Baboon (Figure 4).

This Hellenistic or Roman Egyptian stone statuette depicts a baboon in a traditional squatting pose, measuring 112mm high. Its recorded provenance includes a December 1957 Sotheby’s auction where it was acquired by Ehrenberg, marking it as a significant collector’s piece. However, its history prior to this auction remains entirely unknown. There are no surviving records to indicate where it was discovered or housed before entering the art market.

Because of its intricate physical geometry, the statuette required a highly sophisticated capture strategy. To combat the reflection issues we encountered with the black sherd, we deployed black umbrellas around our lighting setup. This absorbed excess, bouncing light and diffused the light source, dramatically reducing specular highlights on the stone’s surface.

Furthermore, the squatting posture of the baboon created deep shadows and hidden geometries, particularly beneath the chin, around the neck, and at the very base of the statuette. To capture these occluded areas, we adjusted our tripod to shoot from low, upward-facing angles. Just as with the sherds, we photographed the baboon in multiple passes, flipping the object carefully to capture its underside, and successfully merged the resulting chunks into a highly detailed and cohesive 3D model. The finished model can be viewed, rotated, and explored interactively on Sketchfab.

30 Sitting baboon by ICS Ehrenberg Collection on Sketchfab

Reflections

This two-week placement was a profoundly enriching experience. While both Kei and I entered the MakerSpace with some prior digitization experience, working directly with the Ehrenberg Bequest pushed us to explore new possibilities and refine our technical troubleshooting skills.

We are immensely grateful to the Institute of Classical Studies and the Digital Humanities Research Hub for providing us with this opportunity. The environment afforded us a remarkable degree of creative freedom, time, and space to experiment, fail, learn, and ultimately succeed. We gained invaluable practical insight into the complexities of cultural heritage imaging, and we take great pride in knowing our 3D models will help preserve these antiquities for future generations of students and scholars to study, no matter where they are in the world.

To cite this article:

Chen, Kiwa and Kei Dai. 2026. “Digitizing Antiquities: A 3D Photogrammetry Placement at the Institute of Classical Studies.” Stoa Review. Available: https://blog.stoa.org/archives/4420. DOI: https://doi.org/10.83330/stoareview.4420.

Posted in report, Teaching | Tagged 3d imaging, Ehrenberg Collection, photogrammetry, UCL Digital Humanities placement | Leave a comment

Report on Linked Pasts 11 activity: Promoting Collaboration, Linking Communities

Posted on 2026-05-26 by Gabriel Bodard

Linked Pasts 11 Activity – Promoting Collaboration, Linking Communities

Building Pelagios connections through the Digital Classicist Wiki

In this Linked Pasts 11 activity, led by Tom Gheldof (CLARIAH-VL+, KU Leuven) and Sarah Middle (Archaeology Data Service, University of York), we aimed to increase the visibility of Pelagios Partners in the Digital Classicist Wiki, by hosting two themed editing sprints and encouraging further asynchronous activity.

The Digital Classicist Wiki (DC Wiki) is a community-authored resource describing digital projects, organisations and tools, relating to the global ancient world. Pages in the Wiki are grouped by thematic categories to facilitate discoverability and serendipity, as well as forging connections. Contributions are made by volunteers on an ongoing basis, with monthly topic-based sprints being a particular catalyst for activity. While many Pelagios Network Partners already appear in the DC Wiki, some pages are out of date and, prior to our Linked Pasts activity, there were few links in the Wiki between them and the Pelagios Network itself.

In this sprint, we aimed to improve this situation by:

Enhancing existing content: updating pages with current information and applying the ‘Pelagios Network members’ category to create a means of easily grouping and accessing all Pelagios Partner entries;

Adding new content: updating the existing ‘To do list’ to highlight Pelagios Partners who do not yet have a DC Wiki page and prioritising them for creation.

This was also a good opportunity to edit existing entries for Pelagios Partners, to ensure that the information held about them in the DC Wiki was current and accurate.

For those new to the Wiki, we provided links to tutorials (including our newly created tutorial) and documentation, as well as providing support during two synchronous sessions, to optimise coverage for multiple timezones:

Monday 1 December, 3.30-5pm GMT
Monday 8 December, 8.30-10am GMT

During the sessions themselves, 11 participants joined to carry out the above tasks in a friendly and relaxed environment, where they were able to ask questions and learn more about the DC Wiki. Others took part asynchronously, resulting in lively email discussion throughout the duration of Linked Pasts, and a total of 13 new DC Wiki accounts created during the event. In addition to attracting new members, our event also revitalised interest among existing members, prompting them to revisit their content. Together, we made updates to 12 existing DC Wiki pages and added 41 new ones, leading to a more cohesive representation of the Pelagios Network.

As well as enhancing content on the Wiki, participants also enjoyed the event, leading to the following comments that demonstrate enthusiasm for both this and future sprints:

“Thank you for making me access to the DC Wiki, so that I could add our project page. It is really great initiative and I am happy I could contribute at least a little… Please, let me know if you plan to do more events like this in the future, I’d like to attend.”

“Thank you very much for the info and the links… I am glad to participate!”

“Thank you very much for this. We’ll be very interested in participating in this, especially since we are very keen on contributing to the DC Wiki and, moreover, we do not have a Wiki page for the project yet (but we aim to do so).”

Holding a sprint for so many new DC Wiki members, as well as those who had not contributed in a while and needed a refresher, meant that we were able to evaluate our new tutorial. Some of the questions that were asked during the event, as well as a review of the new content that was added, have given us a list of possible changes we could make (with the editorial board) to ensure that all the information is clear and accessible to new members.

This activity not only enriched the quality and connectivity of the DC Wiki, but showcased the diverse projects and institutions affiliated to Pelagios, making their information more discoverable, reliable and current. We additionally hope that this activity has resulted in the extension of the DC Wiki community and has inspired participants to attend future sprints.

We’re planning on holding a similar sprint – this time on enhancing the ‘Linked open data’ category – in mid-July, so please join the Digital Classicist mailing list if you’d like to hear more nearer the time.

Posted in Digital Classicist Wiki, Linked Pasts, report | Tagged linked pasts symposium, pelagios network | Leave a comment

Report on Roman Society Heritage Placement in Institute of Classical Studies

Posted on 2026-05-08 by Gabriel Bodard

Heritage Placement Report: 3D imaging in the Institute of Classical Studies

Guest post by Nori Tang, Durham University

Last summer I joined a two-week placement at the Institute of Classical Studies (ICS) as part of the Roman Society’s Museum and Heritage Summer Placement Scheme. Under Dr. Gabriel Bodard’s supervision, I worked on photogrammetry and Reflectance Transformation Imaging (RTI) of objects from the Ehrenberg Collection. The aim was to create clear and accessible digital models of the Ehrenberg Collection and upload them to Sketchfab, continuing earlier work with 3D photogrammetry.

The Ehrenberg Collection was left to the ICS by Dr. Victor Ehrenberg in 1976. It contains over 150 small antiquities, mainly Greek and Roman ceramics, ranging from the Minoan to Byzantine periods. These objects have been documented in different ways: handwritten catalogues with sketches, typewritten descriptions with photographs, and, more recently, 3D models made through photogrammetry. 3D models allow users to view an object from all angles, share it online, and even 3D print (more or less) accurate replicas, making the collection more accessible for study and teaching.

Before this placement, I had gained some hands-on experience with 3D photogrammetry and RTI as part of my Data Science (Digital Humanities) MSc at Durham University. Most of this work was outdoors, such as recording gravestones in a local graveyard project. While these activities gave me a good understanding of the basic workflow, my practice had been occasional rather than regular.

During the placement, I worked on 3D photographing of a wide range of objects from the Ehrenberg Collection in the Senate House MakerSpace. Each piece came with its own challenges, which led me to practise different aspects of photogrammetry and problem-solving.

The first object was a Gandharan Bodhisattva head (catalogue no. 66), which I worked on together with Aimée Wilkinson (the other student on the SPRS placement, whose report is here). The modelling process went smoothly and finally we decided to take the experiment further by 3D printing and turning it into a fridge magnet. This was a good way to see how cultural heritage objects could be adapted for creative outreach. Another object was an Italian Kylix (catalogue no. 68) — a small kylix with black slip and red painted decoration of tongues and leaves. The surface had a slight reflective quality, and its size made it difficult to handle, especially with its two small handles. I also worked on a Sherd (catalogue no. 47.xii) — a light tan pottery fragment painted black on both sides. The sherd was both small and thin, with a reflective painted surface. Controlling the lighting was especially important here to avoid overexposure on the black paint while keeping the incised decoration visible.

Figure 1: 3D printed magnet of Gandharan Bodhisatva head.

Figure 2: The slightly reflective Kylix, with the shining part appearing hollow in Agisoft Metashape.

The most challenging object was a bronze coin (catalogue no. 78), which was difficult to focus on and required occasional manual adjustments as the turntable rotated. At first, I tried fixing the coin between two pieces of adhesive putty. However, as the turntable rotated, the coin gradually tilted, making the photographs inconsistent. To solve this, I made a simple support device by cutting a small slit into a piece of foam and placing the coin inside. This worked well to stabilize the coin for capturing the upper half of the model, with the lens set to a focal length of f/20 and the ISO at 400. Unfortunately, the underside of the coin was less successful. When I tried merging the two halves, the alignment was poor, and the final 3D model was not satisfactory. At this point, I decided to switch to RTI, which we also practiced this week, and which is particularly good for capturing fine surface details and inscriptions on flat objects like coins. RTI revealed surface patterns that were hard to see in normal photographs, showing its value as a complementary method to photogrammetry.

Figure 3: Data collection. a) Coin supported in a DIY foam stand.

Figure 3: Data collection. b) Positions of the angles used to capture the upper part of the coin.

Figure 4: snapshot of RTI of the coin, showing fine surface details.

Comparing photogrammetry and RTI shows that each method has its strengths. Photogrammetry works best for making full 3D models of objects with complex shapes, while RTI is especially good at capturing subtle surface details, particularly on flat, reflective, or worn items. Although RTI doesn’t produce a full 3D scan, it makes surface textures and relief much easier to see than conventional photos. This helps reveal marks, inscriptions, and indentations, making objects easier to read, study scientifically, and present in an engaging way. Since the Ehrenberg Collection includes other coins, RTI could be a valuable tool for documenting them in the future.

By working on such a variety of objects, I gained a deeper understanding of the technical adjustments needed for different materials and shapes. I also became more confident in adapting my approach when the first attempt did not work as planned.

I am grateful to the Roman Society for supporting this placement. And I would like to express my sincere thanks to Gabby for his patient guidance and encouragement throughout the project. In addition, I am thankful to everyone at the Institute of Classical Studies for creating such a welcoming and supportive environment, and I especially appreciated Katie Boulton the institute manager inviting us to the ICS Library Staff end-of-year party, as their kindness made it easy for me to settle in and feel part of the team.

Posted in report, Teaching | Tagged 3d imaging, cultural heritage, photogrammetry, RTI, society for the promotion of roman studies | Leave a comment

Digital Approaches to Sacred Space in Roman Britain (London, 19th May 2026)

Posted on 2026-05-08 by Gabriel Bodard

Digital Approaches to Sacred Space in Roman Britain

Tuesday 19th May 2026, 10:00-17:30
Institute of Classical Studies, University of London

Registration (in person or online)

Description

Temples and shrines in Roman Britain have been examined since the early modern period, primarily for their architecture and monumental form. Early studies focused on plans, typologies, and building techniques, establishing categories that continue to guide our interpretation. In recent decades, scholarship has expanded to situate these sites within their local and regional landscapes, exploring their connections to roads, settlements, villas. Attention has also shifted toward the material culture associated with these sacred spaces—votive deposits, inscriptions and ritual objects—emphasizing the practices and communities that produced them. More recently, digital approaches have begun to transform the study of Roman sacred space, offering new ways to visualize, connect and analyze archaeological evidence.

This one-day workshop will explore how digital tools and methods help change the way we study and analyse temples and shrines in Roman Britain. By integrating GIS-based spatial analysis, 3D modeling and Linked Open Data infrastructures, the event examines how researchers are using new digital techniques to cast a new analytical lens on sacred sites, but also their material culture and landscape. By doing so, it demonstrates how emerging technologies create new opportunities to test existing interpretations, question established models, and formulate new hypotheses about sacred space in Roman Britain.

Organised by Maxime Guénette (University of Montréal), John Pearce (King’s College London), Gabriel Bodard (University of London).

Programme

10:00-10:30 Coffee and registration
10:30-10:45 Welcome remarks
10:45-11:15 Fragmented Data and New Networks in the Study of Roman Religion
11:15-11:45 Mapping Sacred Spaces in Roman Britain: A New Corpus on Wikidata
11:45-12:15 RomAniDat and the Pandora platform: using spatial modelling software for analysing animal bone data from temple sites
12:15-12:30 Discussion
12:30-14:00 Lunch
14:00:14:30 Digital approaches to space and excavation at the Roman shrine landscape of Teffont, Wiltshire, UK
14:30-15:00 Spectres of the Past: Haunting, Uncertainty, and Space for Digital Imaginaries
15:00-15:30 Integrating landscape-scale prospection, high-definition 3D GPR and targeted excavation at a newly discovered Romano-Celtic temple in Somerset
15:30-16:00 Coffee
16:00-16:30 Identifying sacred sites in the rural landscape
16:30-17:00 Londinium AD 215: imagining urban sacred space in Virtual Reality
17:00-17:30 Closing remarks and discussion

Abstracts

Fragmented Data and New Networks in the Study of Roman Religion

Speakers: Alessandra Esposito (King’s College London), Francesca Mazzilli (Università degli Studi di Bari Aldo Moro)

This contribution examines the growing ‘digital turn’ in the study of ancient religions, with a focus on the Roman world, highlighting both methodological advances and persistent challenges. It critically addresses the fragmented landscape of current digital approaches, where existing projects (often institutional, geographically bounded, and unevenly cited) remain dispersed, reflecting disciplinary and regional isolation.

Alongside case studies from the Near East and Dacia, the inclusion of examples from Roman Britain highlights the richness of regional research traditions. These cases demonstrate how digital tools (GIS, network analysis, relational databases) illuminate patterns of mobility, cult transmission, and material culture, including processes of use and reuse.

Despite these resources, the field lacks unified standards, interoperable repositories, and comprehensive infrastructures beyond isolated initiatives, with some exceptions. The article calls for FAIR data, improved searchability, and clearer definitions of ‘digital religion’ for both specialists and newcomers. It advocates for cross-sectoral collaboration on frameworks and new research networks (ideally via a COST Actions) to overcome fragmentation, enhance public engagement, and address urgent heritage challenges.

Mapping Sacred Spaces in Roman Britain: A New Corpus on Wikidata

Speaker: Maxime Guénette (Université de Montréal)

This presentation introduces a new, comprehensive corpus of nearly 500 temples and sanctuaries identified across Roman Britain, offering the most up-to-date systematic inventory of sacred spaces from this period. Drawing on recent archaeological scholarship, the corpus integrates sites ranging from major Romano-Celtic temples to rural shrines and military cult sites, reflecting the rich religious diversity of the province.

A key methodological contribution of this work lies in its use of Wikidata as a platform for producing Linked Open Data (LOD). By structuring archaeological data within Wikidata’s open, collaborative framework, the corpus becomes fully interoperable with other digital humanities and heritage datasets, enabling cross-referencing with epigraphic, numismatic, and spatial databases. This approach lowers the technical barriers typically associated with LOD publication, making the data accessible to both researchers and the wider public. Finally, we also reflect on the limitations and criticisms associated with Wikidata as a publication platform for historical and archaeological datasets, situating them within broader debates on data quality, editorial control, and long-term sustainability.

RomAniDat and the Pandora platform: using spatial modelling software for analysing animal bone data from temple sites

Speakers: Tony King (University of Winchester), Dominika Schmidtova (Masaryk University), Ricardo Fernandes (Max Planck Institute for Geoanthropology)

This presentation describes the RomAniDat project for analysing datasets of animal bones across the Roman world, and the project’s use of Pandora software as developed by the Max Planck Institute for Geoanthropology. The software enables user-entered data to be analysed for spatial modelling and mapping, as demonstrated by results for Roman Italy, Britain, the Netherlands and the Empire as a whole. Application of this software platform for animal bone (and other) data from religious sites in Britain is modelled and discussed.

Digital approaches to space and excavation at the Roman shrine landscape of Teffont, Wiltshire, UK

Speaker: David Robert (Cardiff University)

Excavations at Teffont have revealed a complex multi-period landscape centred on a ridgetop sacred space of predominantly Roman date. GIS has been critical to the project, allowing the integration of multiple forms of data, and new insights to emerge from the site. The paper will explore how kernel density mapping of multiple strands of non-invasive data, alongside integration of aerial, geophysical and excavation data, has enabled us to better understand the site, which is within woodland, and therefore has significant challenges to understanding the below ground archaeology. The paper will further reflect on the accuracy and value of the methodologies chosen to explore the site.

Spectres of the Past: Haunting, Uncertainty, and Space for Digital Imaginaries

Speaker: Anna Collar (University of Southampton)

Everywhere, we live alongside the material remains of monuments, but although the terms ‘reuse’ and ‘afterlife’ are commonly-used to describe what happens with or to monuments beyond the time of their primary use, these terms carry serious conceptual issues: implying both passivity in monuments’ roles in later periods, and imposing hierarchies of significance on our interpretation of them. The Avebury landscape as a ‘heritage asset’ is a case in point: the monument has been ‘preserved’ as a composite ‘mythical moment’ of Neolithic and Bronze Age—and the significance of the monuments in later periods is ignored (Hughes 2021: 9).

Ruination and abandonment may lead to memories and meanings associated with monuments being forgotten or disrupted, but this does not remove the agency of these places. Instead, ruination makes monuments sites for haunting: that is, places that may be poorly understood and which may affect the present in unexpected ways (Derrida 1994). Such places of uncertainty are places where ‘spirits thrive […] in conditions of doubt rather than belief.’ (Bubandt 2017: 125). Haunting radically alters our conception of how earlier monuments impact later peoples’ lives, experiences, and cosmologies by acknowledging the agency of monuments as loci of uncertainty. I suggest that the monuments at Avebury became such places of doubt in the Iron Age-Roman period: spiritually dangerous, and in need of continuing boundary-work, exemplified by a sanctuary, constructed beyond the Winterbourne but with an eye on the henge. I will finish by thinking about the role that digital technologies such as VR and AR can have in helping us to visualise and imagine parts of the past that are forgotten and overlooked.

Integrating landscape-scale prospection, high-definition 3D GPR and targeted excavation at a newly discovered Romano-Celtic temple in Somerset

Speakers: Jodie Lewis (University of Bradford), Michael Pisz (University of Bradford)

Recent archaeological investigations on the Mendip Hills, Somerset, have revealed a complex multi-period landscape extending from prehistory to the post‑medieval period. Among the most significant findings is a previously unknown Romano-Celtic temple, located within a notable concentration of prehistoric monuments, offering new insights into long‑term ritual activity in this upland environment. This paper presents the results of two field seasons focused on the temple and the surrounding area. Remote sensing data, spatial analysis, and large-scale magnetometer survey provided a landscape-scale assessment and identified the wider archaeological setting of the temple complex. This broad prospection was followed by a high-definition 3D ground-penetrating radar survey, which enabled detailed characterisation of the subsurface remains and revealed the complexity of the temple at a much finer scale. These geophysical results were then tested through targeted excavation, creating an integrated workflow from landscape screening to detailed interpretation and ground-truthing. The results demonstrate how state‑of‑the‑art digital datasets can be integrated with excavation records to produce a more comprehensive reconstruction of buried features and their wider landscape context.

Identifying sacred sites in the rural landscape

Speaker: Stephen Clifton (University of Kent)

Are we sure that we understand what is happening in the rural areas of Roman Britain away from the urban centres and the military? There has been a tendency to characterise those complex Romano-British sites known as ‘villas’ along continental lines. Imagining a grand domestic establishment incorporating farming practices. Yet when we look more closely, many of the elements that we would expect to see are not present and there are other elements that we find difficult to explain using our current model. Digital modelling can be a useful tool to help us to reimagine these spaces in order to see how they would have been used so that we can better test out our models and theories. Using the well preserved site at East Farleigh in Kent, long characterised as a Roman Villa, we can see how the sacred rituals emerge from a reinterpretation of the evidence.

Londinium AD 215: imagining urban sacred space in Virtual Reality

Speakers: Dominic Perring (University College London), Jake Nixon (Historic VR)

A virtual reality experience, Londinium AD 215, will shortly launch on Steam. Over three years in the making, this offers a free-roaming VR and desktop experience across a digital reconstruction of an entire Roman city. It draws faithfully on the evidence of hundreds of geo-referenced archaeological investigations, but it still takes a considerable leap of imagination to rise above the fragmentary remains uncovered on City construction sites. The exercise has required us to overcome the limitations of both digital technologies and archaeological evidence to build credible urban landscapes and, in turn, to better visualise and understand the subject of our archaeological research. The ritual and sacred properties of the landscape are key to this reconstruction exercise. Not just in giving life to temples and shrines, including some whose existence is largely conjectural, but also because the city itself was designed as a theatre for religious and civic ritual. The need to recreate absolutely everything that might have been visible in Roman London has required us to develop new arguments and ideas about how the city functioned. This presentation – which will use case-studies drawn from London’s temple districts – will be our first opportunity to consider some of the lessons learned.

Posted in Conferences, Events | Tagged GIS, networks, sacred space, spatial humanities, spatial technologies | Leave a comment

Report on LP11 activity: Epigraphic Data in Wikidata

Posted on 2026-05-05 by Gabriel Bodard

Linked Pasts 11 – Epigraphic Data in Wikidata: From Ontology to Practice

Dates: 2–3 December 2025
Duration: 1h30 each day
Convenors: Anna Clara Maniero Azzolini, Maxime Guénette, Anne Chen, Emily Helm
Format: Online
How participants joined: SAS event (LP11 programme)
Type of audience: Independent Scholars, Researchers, Professors, students (Historians)
Requirements: Wikidata account
Objectives: This session aimed to introduce and explore the use of Wikidata, a collaborative knowledge base, for epigraphy, providing both theoretical knowledge on linked open data (LOD) and ontologies, and practical skills for entering, querying, and managing epigraphic inscriptions.

Report

Day 1
Introduction, Theoretical Framework and Hands-On Practice

The first session opened with a short survey designed to understand participants’ academic backgrounds, their previous experience with Wikidata, and their motivations for attending the workshop. Although all participants had created a Wikidata account in advance, as requested, most had not yet engaged in structured editing within the platform. Some were particularly interested because of their work with Linked Open Data (LOD), while others had prior experience editing Wikipedia, the sister project of Wikidata.

Theoretical Framework: Linked Open Data and Wikidata

Before discussing Wikidata itself, we examined the concept of Linked Open Data (LOD) as a model for publishing structured data online in a way that is open, interoperable, and machine-readable. Rather than allowing data to remain in isolated silos, LOD promotes standardisation and connectivity, enabling datasets to be linked across institutions, disciplines, and national boundaries.

Within epigraphy, several highly valuable databases already exist (such as EDR, EDH, EDCS, EDB, and RIB) offering extensive documentation of inscriptions, including texts, editions, metadata, and findspots. While these resources are indispensable, they also present certain limitations: uneven geographical or chronological coverage, inconsistent multilingual support, restricted possibilities for user contributions, and querying systems that, although useful, can only operate within the boundaries of their individual databases.

What is often lacking is a more centralised, multilingual, and collaborative space in which inscriptions can be compared, corrected, enriched, and queried flexibly across datasets. In this context, Wikidata becomes particularly significant: as a collaborative knowledge base within the Wikimedia ecosystem, Wikidata stores structured data (rather than narrative articles as Wikipedia). Its infrastructure is based on Wikibase, an open-source software that both institutions and individuals can install.

An example discussed during the session was the way Wikipedia infoboxes automatically retrieve information from Wikidata: a single correction to a data item can propagate across multiple language versions of Wikipedia, ensuring consistency and broad dissemination: this demonstrates the scale and impact of structured data within a global knowledge network.

From a research perspective, Wikidata’s appeal lies in its openness, scale (with nearly 120 million entities), and full compatibility with the Linked Open Data model. In the epigraphic domain, inscriptions can be linked to related datasets such as Trismegistos, Pleiades, RIB, or museum catalogues, thereby situating them within wider archaeological, prosopographical, geographical, and material contexts. Given the inherently interdisciplinary nature of epigraphy, this interconnected model is particularly advantageous.

Data Structure and Modelling

A key component of the introduction involved explaining how Wikidata structures information. All data are organised in the form of triples (subject, predicate, object) comparable to concise declarative statements. Each element is assigned a unique identifier: items are labelled with Q-numbers and properties with P-numbers. In this way we can ensure clarity and disambiguation, allowing both humans and machines to interpret statements precisely.

Participants were also introduced to qualifiers and references. Qualifiers allow statements to be contextualised (for instance, specifying dates, certainty levels, or types), while references ensure verifiability (which is crucial for scientific data).

Practical Session

Following the theoretical overview, participants moved into breakout groups for a practical exercise. Anna Clara led the editing group, while Maxime and Emily guided querying activities.

The editing exercise focused on manual editing through the Wikidata interface, enabling participants to become familiar with adding statements, identifiers, references, and qualifiers. This slower, hands-on approach allowed for careful reflection on modelling choices and provided space for real-time feedback and discussion.

Meanwhile, the querying group explored how structured data can be retrieved and analysed using SPARQL. Participants were introduced to the Query Builder as an accessible entry point, followed by a brief demonstration of the more advanced Query Service for complex searches. The aim was to show how modelling decisions directly affect the kinds of questions that can be asked of the data.

Day 2
Advanced Modelling, Validation, and Case Studies

The second session began with a recap of the editing principles introduced in the first session, followed by more advanced demonstrations of query construction and exploration of additional Wikidata services in break-out rooms.

Epigraphic WikiProjects and Data Models

The session highlighted two epigraphy-related initiatives within Wikidata: the IDEA project and the Altinum project, both led by the authors. These projects provide concrete models for representing inscriptions and serve as collaborative spaces in which contributors discuss standards and best practices, thereby offering participants in the working groups the opportunity to engage directly with established data models.

The IDEA Project reunites material from Dura-Europos in Wikidata and by modelling inscriptions, papyri, coins, buildings, and other artefacts, the project demonstrates how LOD can reconstruct archaeological and historical contexts through highly queryable relationships. The Altinum Project was presented as a case study of integrating data from EDR, EDCS, and unpublished theses into Wikidata. It illustrated how consistent modelling enables inscriptions and associated individuals to be embedded within a shared, multilingual ecosystem, enhancing visibility and interoperability.

Anne and Emily also presented current epigraphic data models in Wikidata, explaining how inscriptions, metadata, and relationships to places, people, and objects are encoded. They discussed strategies for verifying whether inscriptions already exist in the database and introduced ShEx schemas as a method for validating data structures and ensuring modelling consistency.

Breakout Sessions

After this theoretical introduction, participants joined the break-out rooms according to which topic they wanted to explore. Anna Clara led the editing group, Anne the ShEx Schemas and Emily the querying group. Participants engaged with practical tasks tailored to each thematic focus, deepening their understanding of the topics.

Outcomes and Reflections

By the end of the two-day workshop, participants had:

Developed a clear understanding of Wikidata as a collaborative, structured knowledge base aligned with the principles of Linked Open Data.
Gained practical experience in manual editing, including the use of properties, items, values, qualifiers, and references.
Acquired foundational skills in querying data using SPARQL, Query Builder, and, to a basic level, the Query Service.
Developed awareness of current epigraphic data models and the importance of consistency and validation (including ShEx schemas).
Reflected on the broader challenges and opportunities involved in modelling inscriptions within a global, open ecosystem.

Overall, the workshop successfully combined discussions with practical applications. It demonstrated the potential of Wikidata to support digital epigraphy by fostering collaboration, enhancing interoperability, and increasing the discoverability and reusability of inscriptional data across projects and disciplines.

The number of questions raised throughout the sessions, together with the consistently positive feedback, reflected participants’ engagement and willingness to contribute actively; moreover, several inscriptions were modelled with notable accuracy and coherence, demonstrating a solid understanding of the principles discussed.

Posted in EpiDoc, Linked Pasts, report | Tagged epigraphy, linked open data, linked pasts symposium, open data, Wiki, wikidata | Leave a comment

A Feasibility Study for an EFES publication of the Inscriptions of Roman Pollentia, Mallorca (Balearic Islands, Spain)

Posted on 2026-04-29 by Gabriel Bodard

by Sarah Williams (University of London)

(Note: to see my EpiDoc files and EFES customisation, and for local links in this post to work, you will need to download my fork of EFES from https://github.com/Minna-Loushe/EFES-Pollentia, install and run from the command line, as per the instructions at https://github.com/EpiDoc/EFES/wiki/Installation, and discussed below.)

I. Project Outline and Aim

This project was inspired by the EpiDoc/EFES week of SunoikisisDC 2024-25 and by I. Vagionakis’ 2021 article about her Cretan EpiDoc/EFES database. EpiDoc is an international, collaborative effort that provides guidelines and tools for encoding scholarly and educational editions of ancient documents (described in detail at https://epidoc.stoa.org/). EFES (EpiDoc Front End Services) is a custom and readily customizable platform for publication and search/indexing of EpiDoc files (see https://github.com/EpiDoc/EFES).

The EFES user guide states that a student with a working knowledge of EpiDoc should be able to use EFES without prior experience of computer programming or software development, criteria I fulfil. My aim was to explore the feasibility of getting EFES up and running ‘out-of-the-box’ for the inscriptions of the Roman town of Pollentia on Mallorca; I aimed to utilise any existing open access digital resources together with a hard copy of the CIBalear corpus as published in 1965 (CIBalear = Veny, C. Corpus de las Inscripciones Balearicas Hasta la Dominacion Arabe). The corpus of inscriptions from Pollentia numbers around 80, of which Veny knew 44 in 1965; Pollentia remains an active archaeological dig site with discoveries subsequent to 1965 regularly published, making it ideal in size and scope for this project.

[Since 12 May 2025 CIBalear has become freely downloadable online, a wonderful work of scholarship, full of stories of antiquarian discovery, forgery, thefts, the Spanish civil war and Castilian/Catalan politics.]

II. Methodology

There are two Roman Pollentias—one in Mallorca and one in Liguria, Italy. My first task was to discover any existing digital editions of the relevant inscriptions. To find the Mallorcan results in the Epigraphik-Datenbank Clauss/Slaby (EDCS) you must first search by “Pollentia”, scrape your results, and then by “Inca”; the inscriptions are quirkily categorised to these different towns for no apparent reason. Starting with the EDCS results and using the direct links on that site to the Epigraphic Database Heidelberg (EDH) is the only successful way of finding the inscriptions in EDH (for reasons explained below). EDH contains only inscriptions published in Veny 1965 (why?) although CIBalear54 is missing—despite being neither more fragmentary nor more likely a fake than any other. EDCS contains the entire corpus, but only EDH provides downloadable EpiDoc files.

To begin my project I forked EFES to my GitHub account with GitHub Desktop running (following instructions at https://github.com/EpiDoc/EFES/wiki/Installation) and watched M. Filosa et al’s completely indispensable 2021 Epidoc/EFES online training week. I then imported EpiDoc files from EDH into EFES and named them by their CIBalear** references. Extremely naively, I expected EFES functionality would work off the mark-up (which I knew existed in my imported files), and that I could move on to updating my EFES corpus with new EpiDoc files for inscriptions discovered since 1965. I now realise how hopelessly unrealistic this was and, in my next section, will document my problem-solving for all the functions that EFES offers one by one. My aim in showing the (sometimes minute) details of my trials and errors is to attempt to be of use to anyone else beginning an EFES project for the first time.

I note in passing that there is a simpler way to get EFES running (especially in the absence of a Github account) by downloading a ZIP file. Full instructions are also included on the Installation page.

III. Exploring EFES out-of-the-box

III.1. Indexing

The onomastics and prosopography of Pollentia are particularly fascinating so I decided to try and get the indexing of ‘persons’ or ‘personal names’ working using EFES’ authority lists. It turned out to be my most challenging task—here is what I learned:

EFES ‘out-of-the-box’ has no authority list for ‘persons’ or ‘personal names’ but I understood how to make one (see M. Filosa et al 2021—especially Day 4/Part 1). I followed the instructions and spent a long day completely failing to index a single person or name. Eventually I worked out why; EDH EpiDoc files are marked up for persons in an unexpected way. For example, in CIBalear30 the edition is as follows:

Cornelius Atticus is then also marked up as a person in a separate section in the .xml (an equivalent function to the use of authority lists in EFES, designed to remove the need for such tagging in the text itself), as follows:

This appears on the EDH website as follows:

But in EFES epidoc-persons-index-to-solr.xsl only looks within the edition:

I can see that there are solutions. One could change the template matching in the .xsl file—beyond my capability. Or one could change the mark-up within the editions, which is within my capability but was not possible in my available timeframe.

In my quest to get an index working I moved on to a feature where I knew that the mark-up was as I expected. EFES comes with an index for ‘places’ which then appear as ‘toponyms’ and an authority list for ‘mentionedplace’. I marked up mentioned places and updated the authority list—it worked.

These are all the instances of mentioned places within my mini corpus. And I was able to solve the problem that out-of-the-box EFES does this…:

… due to having watched the online training (M. Filosa et al 2021, specifically Day 4/Part 2, approx. mins 19-23). I have mentioned that these sessions are indispensable.

III.2. Facets and Searching

A satisfying side effect of sorting out the indexing for mentioned places is that the faceted search function works with no further intervention:

This is the correct information and you can click on each item and it returns the correct list of inscriptions.

I knew from the User Guide that only ‘some’ of the pre-existing facets require the use of authority lists so I worked my way through. For example, without my doing anything the search for ‘found provenance’ returned this….

…but these are not clickable and you cannot discover which 11 or 43 inscriptions belong to each category. Unfortunately, almost all (all?) the ‘place’ related information in these EDH files is incorrect. geonames.org/2512432 is modern Pollença, a town in Mallorca, but not particularly near Pollentia, which is in modern Alcúdia. geonames.org/2510769 is, simply, Spain and should therefore apply to all these inscriptions. Pleiades:383745 (found throughout) is, unfortunately, the site of Pollentia in Liguria, Italy. (The ‘current repository’ data is also incorrect as the Mallorcan authorities have moved the inscriptions around.) Searching the pre-existing facets does therefore have the side benefit of being a fast method to check the validity of any imported mark-up.

I moved on to ‘origin place’, filled in the authority list and got it working:

This ‘Pollentia’ is now clickable and will return CIBalear20, which is the only one I marked up. All the rest would be the same as we have little granularity beyond simply the site name for the largely antiquarian discoveries.

Similarly with ‘support material’:

The first two items are not yet clickable but for the third, ivory, I have marked up the authority list (retaining the eagle-network.eu information in the authority list).

I do not fully understand why some facets are picking up data directly from the EpiDoc files and returning partial (although not useful) results, but neither does one need to understand this to get the functionality working. I note that once I have removed any informational URL into the authority list the output no longer appears in blue in the inscription .html (words in blue are clickable). The word search box worked nicely without my doing anything. The editions are not lemmatised but that is not in the scope of this project.

III.3. Date Slider

Initially the Date Slider was not working. I found that all the dates are marked up:

<origDate notBefore-custom=”0001” notAfter-custom=”0051” datingMethod=”http://en.wikipedia.org/wiki/Julian_calendar“>1 AD – 51 AD</origDate>

One gets faster in EFES with practice so I knew (without resorting to tutorials!) that it would not like the ‘-custom’—with it removed the Date Slider worked perfectly.

III.4. Adding Images

This is straightforward once one has been intensely frustrated and then noticed that .jpg is very different to .jpeg in EFES.

III.5. Multilingualism

It is self-evident that EFES-Pollentia should at least be fully bilingual. Although the first language of Mallorca is Mallorquin Catalan, Veny published in Castilian, as do many other scholars of Pollentia. Perhaps therefore it should be fully trilingual and I explored the EFES functionality for this.

The EDH files contain no Translation or Commentary (or Apparatus Criticus). However, I had Veny’s so, using CIBalear20 as my test case, I marked up the translation as follows (the Castilian is Veny and the English is my translation of Veny):

This appears in the .html as follows:

I followed the same process for the Commentary. Only the English then appears in the .html but to see the Castilian you can change ‘en’ to ‘es’ at http://127.0.0.1:9999/en/inscriptions/CIBalear20.html. (Note that links to pages within my project will only work if you have EFES running locally.) To fully customise the languages of a project in EFES, as the User Guide points out, one would need to know how to write .xsl and .xmap files.

III.6. Bibliographic Concordance

I had underestimated the ‘concordance’ element of this function. Examining it earlier would have helped me to think about how to name files and organise this project. I proved the concept of using this function by adding EDH to my CIBalear20 file but it would be a lengthy task to complete the full concordance.

The first <listBibl> is as entered by the EDH editor and the second by me using the authority list. The output is as below:

III.7. EFES-Pollentia Homepage

To create a homepage I imitated examples I had of existing EFES projects along with some trial and error. To innovate one would need to learn CSS for this (not covered in the EFES online training).

Note that the homepage for Vagionakis’ wonderful Cretan project (available at https://github.com/IreneVagionakis/CretanInscriptions, with Github or ZIP instructions for access) looks different depending which browser you use—the first screenshot is Safari and the second Firefox:

(1)

(2)

IV. Discussion/Conclusion

My results can be found at: https://github.com/Minna-Loushe/EFES-Pollentia (as for all EFES projects one needs EFES running, as detailed in the Installation pages, to explore this in detail).

I was able to get this EFES project up and running ‘out-of-the-box’; all the functionality works with my sample of files.

I certainly learned the importance of an appropriate ‘edition structure template’; I wonder how best to create this since one knows the most about a corpus when quite a long way through a project? Although I imported EpiDoc files from EDH I wondered if it would have been easier to start from scratch and, for my project, this may remain an open question due to the inaccuracies throughout the imported files. Other project authors may not encounter the same difficulties, and the EDH files are there to be re-used, as discussed by F. Grieshaber in 2019.

EFES’s own suggested uses are both the publication and management of a student’s own workflow, either individually or collaboratively, and I notice that other projects use the functionality to help with this. For example, Dobias-Lalou et al.’s igcyr2 uses a search facet as follows:

I. Polinskaya, A. Ivantchik et al at IOSPE more than meticulously track their changes:

By ‘commenting out’ one can write notes within an EpiDoc file to evidence which areas still need work, or which have been checked. One might want to index/search by the last date a file was changed, for example, or know it is considered finalised—to use the functionality to manage the workflow.

EFES is more than useful for a student working on a corpus: the iterative process creates immediate feedback and the problem solving itself teaches invaluable lessons about one’s work (and EFES somehow starts teaching about itself). In the ‘real world’ one would soon want to innovate and seek to learn XSLT, CSS and Cocoon Sitemaps, hardly a bad thing.

Posted in EpiDoc, Open Source, Prepress, SunoikisisDC | Leave a comment

Digital Classicist London 2026 programme

Posted on 2026-04-24 by Gabriel Bodard

The Digital Classicist London seminars present aspects of the ancient or pre-colonial worlds, including archaeology, cultural heritage, history, language, literature and their reception, and address innovative digital approaches to research, teaching, dissemination or engagement, especially work across or between disciplines, and collaborations with partners outside academia that serve the research agendas of all parties.

Seminars are hosted by the Institute of Classical Studies, University of London, at 17:00 BST on Fridays (except for 19th June, at 13:00 BST). All sessions are streamed live to Youtube, and several are also in-person in London (booking required for in-person attendance).

Fri, 12th June 2026, Sarah Rohde (University of London), Classical Antiquity in Video Games: Combining Visual, Narrative and Interactive Elements (Youtube) (registration)
Fri, 19th June 2026 (13:00 BST), Alina Kozlovski and Sarah Lawrence (University of New England), A Museum of Words and Deeds: Combining artefacts and Latin texts in an online exhibition (Youtube) (online only)
Fri, 26th June 2026, Federico Di Pasqua (University of Zurich), GraphRAG and the Ancient World: Toward a Spatial AI Research Platform (Youtube) (registration)
Fri, 10th July 2026, Michele Corti (University of Siena), Multimedial Edition of Hero Alexandrinus’ Pneumatica (Youtube) (registration)
Fri, 17th July 2026, Burak Boşdurmaz (University of Koç), Mapping the Blue Horizon: A Digital Corpus Approach to the “Blue-Range Color Complex” in Classical Latin Literature (2nd c. BC–AD) (Youtube) (online only)
Fri, 31st July 2026, Jesse Obert (University of South Florida), Warriors without War on Archaic Crete: Testing Paradigms with Big Data and Social Network Analysis (Youtube) (registration)

All welcome.

Posted in seminar | Leave a comment

Report on PN People WG Sample Data Hackathon at LP11

Posted on 2026-03-13 by Gabriel Bodard

Gabriel Bodard, University of London
Jun Ogawa, University of Tokyo

On the 1st and 4th of December 2025 we held two meetings of the Pelagios Network’s People Working Group (which works towards consensus and standards for common representation of historical people in structured linked open data) as a programmed activity of the Linked Pasts Symposium (LP11), following up on a preliminary WG meeting on October 30th.

These online-only workshops were designed to explore some sample person-data (extracts of between 100–1000 records per project in a variety of formats) that were collected in our Github repository at previous meetings of the People WG. We introduced these sample datasets to participants and explained the possibility of cloning or downloading the repository, for those interested in hands-on tinkering with the data, or simply exploring the metadata on-screen via the Github interface, for more conceptual analysis. The goal of these workshops was to explore alignments and interoperability between datasets.

Starting at the first meeting, in October, two working groups approached the question:

(1) Led by Gabriel Bodard with many contributors, especially helpful among whom were Emily Helm and Camillo Pelizzari, the first group looked at the metadata and made note of links or alignments between them, including shared fields/properties, terms or vocabularies, common persons or other entities. This was continued in December with an extremely fruitful discussion between Gabby, Jun and Burcun Aksahin comparing library records on Aeschylus from IDREF and GND with the parallel information in Wikidata.

This activity, while relatively theoretical and “low tech,” is an essential component of both the ongoing, more hands-on data processing exercises, and the underlying task of identifying which fields and properties are common (if not universal) and therefore essential for an interoperability format or recommendation. A common format for all prosopographical and other historical person datasets is not feasible (or desirable), but a recommended minimal subset of data fields to expose in a common format would enable cross-searching, interrelationship and other rich research use of many diverse datasets.

(2) Led by Jun Ogawa, the second group experimented with more hands-on activities, including processes such as converting sample data into Wikidata-style RDF, performing data reconciliation, storing and querying the data in a triplestore. We basically worked individually, Nurdan Atalan Çayırezmez and Jun Ogawa working on transforming non-RDF data into defined RDF and Akihiro Kameda on converting RDF into wikidata format.

Thanks to this session, we could have useful technical experiences on dealing with different formats of person-data, and gained insights into what tools or skills were useful or even necessary for future advancement. These insights will eventually help when thinking about how our final output (recommendations, tools, registries) should be designed and developed to increase interoperability among historical people datasets.

Considering these different approaches, the range of background and expertise expected was quite wide, including participants with a range of backgrounds and expertise were welcome, from those with interest in cataloguing or prosopography to those familiar with more technical aspects of Linked Data, including OpenRefine, triplestores, SPARQL. Initial reports on these meetings were discussed on the Ancient People mailing list, and minutes posted to the LOD People Wiki.

For the next steps, we plan to hold another hands-on, technical session at Linked Pasts Japan II, which will be held in February 2026 in Tokyo (most probably hybrid), to continue transforming data into common RDF format and uploading it to a triplestore to query. We expect many Japanese DH scholars and practitioners who are interested and familiar with technical aspects of Linked Data will join the session to make a great advancement on this work.

Also, we will continue organizing regular meetings in 2026 of the PN People WG on topics related to what we worked on at the symposium including ontology comparison, project reports, or more technical issues.

If you would like to get involved in this activity or working group in any way (membership of the Pelagios Network is not a requirement!), simply sign up to the Ancient People mailing list to be kept up to date on meetings and other events.

Posted in report | Tagged linekd open data, linked pasts symposium, pelagios network, snap:drgn | Leave a comment

Digital Classicist London 2026 call for papers

Posted on 2026-02-18 by Gabriel Bodard

The Digital Classicist London seminar invites proposals for the Summer 2026 series. We are looking for papers on any aspect of the ancient or pre-colonial worlds, including archaeology, cultural heritage, history, language, literature and their reception, and that address innovative digital approaches to research, teaching, dissemination or engagement. We particularly welcome presentations that reflect work across or between disciplines, and collaborations with partners outside academia that serve the research agendas of both/all parties.

Seminars will be held on Fridays in June and July in Senate House, University of London, and will be simultaneously streamed to remote audiences on Youtube, but we hope most speakers will be physically present in London. We have a small budget to support travel for speakers within the UK.

Please send an abstract of 300 words or less to <gabriel.bodard@sas.ac.uk> (clearly marked Digital Classicist London) by the end of Monday 16th March 2026. In the email, please indicate if you would be able to present in person in London or online only.

The organisers:

Gabriel Bodard (ICS)
Maxime Guénette (Montréal)
Elizabeth Koch-Kölük (ICS)
Anna Clara Maniero Azzolini (ICS)
Katharine Shields (KCL)
Sarah Williams (ICS)

Posted in Call for papers | Tagged London | Leave a comment

Review of Ugarit and Translation Alignment using Petronius

Posted on 2025-09-29 by Gabriel Bodard

By Lucy Yang (KCL)

Texts aligned in Ugarit iAligner for this project: Satyricon 132.1, 132.2, 132.3, 132.4

The Satyricon by Petronius enjoys its modern reception as both a social satire to Neronian Rome, as most scholars believe its date to be, and as a pioneering example of the ancient novel. While the Satyricon’s content and form contribute to its fascination, they also complicate its translation, particularly in terms of preserving semiotic and structural integrity (Kay, Computational Linguistics, 1993). In terms of its contents, the Satyricon has had a reputation for its sexual explicitness and subversion rarely found in classical texts from the ‘classical canon’. Its fragmentary state and the nature of its content kept the Satyricon from a rightful examination as an ancient text, before its ‘scholarly worth’ was established by J.P. Sullivan in 1968 (The Satyricon of Petronius). Even then the Satyricon has continued to present interpretive challenges.

On the other hand, translation of the Satyricon to English is inherently a form of interpretation. I argue that translation of the Satyricon is especially important because it engages with the aspects of semiotic integrity and censorship, as demonstrated above, particularly relevant to the Satyricon. By studying translations, we get an insight of how the text is manipulated for an English-language readership. How is the form of the Satyricon mediated through translation, what has been presumed by the translation about the purpose of the text, how do different translations change the reader’s interpretation? Using text alignment to examine different translation intentions, this project evaluates how this digital tool may be applied in translation studies and Classical reception.

As Hardwick notes in her seminal 2000 book on translation theory and reception theory, Translating words, translating cultures, the main function of a translation of ancient texts is to provide ‘a contemporary means of understanding and responding to the ancient work’ (11-12). Translations offer a lens through which we can observe how a text was understood in different periods. The Satyricon was renowned for its ‘obscenity’ more so than its textual genius. This reputation often overshadowed its textual sophistication, leading to censorship through omission or alteration in translation (see Firebaugh 1922, below). The translations chosen for the project are more than 100 years apart, aiming to
demonstrate how the increased scholarly recognition has changed the ways in which the Satyricon is translated.

The two chosen translations are named A and B respectively. Translation A is by W. C. Firebaugh in 1922 (text from WikiSource). Translation B is by Gareth Schmeling, published by Loeb Classical Library in 2020. I have selected the Satyricon 132 for its moderate length to keep the project realisable and the conclusion more specific to textual analysis. Moreover, this passage incorporates both verse and prose—a prominent feature of the Satyricon ‘s form that makes it one of the few ancient novels. Content-wise, the passage includes features that contributed to the Satyricon’s early censorship as well as its later scholarly interest. It portrays both the ‘indecent’ element, Encolpius’ rage and shame over his sexual impotence and the literary aspect of the Satyricon’s intricate engagement with classical literary allusions (Conners, Petronius the Poet, 1998: 145; Rimell, Petronius and the Anatomy of Fiction, 2002). This project aims to demonstrate how digital textual alignment can highlight alterations and adaptations from Latin to English at both the formal and content levels. From evaluating the alignment’s ability to identify and display these differences, it aims to highlight how practicing textual alignment may help non-linguist classicists working mainly with translations to overcome the untranslatability of language barriers.

The methodology with translation alignment consists of the pairing of equivalence between the original text to the translated text(s), on word, phrase or sentence levels (Yousef and Janicke, Survey, 2021). General guidelines can be accessed on the community web site for different language pairings of original/translation. In practice, how and what to align can be project-specific. The guidelines for Latin-English alignment as provided by Valeria Irene Boano for Ugarit are mostly applied here. Most notably, this means aligning both the verb and its subject in English with the corresponding verb in Latin (Boano’s guidelines, 8.2). The Latin text and both translations are put through a single document to clean out the noise including page numbers and annotations, the punctuations are left in the text for an easier manual alignment process but not aligned. To keep a consistency for the alignment rate, definite and indefinite articles (‘a’, ‘an’, ‘the’) are not aligned. Latin verbs that imply the subject are aligned with the subject in English when possible. This project evaluates interpretative integrity in different translations, reflecting the methodology of a non-linguist classicist with limited knowledge of the ancient language. To fulfil these needs, only words with direct semiotic correspondence are aligned with help from Latin-English dictionary.

Graphic simulation makes the process of textual alignment efficiently interactive as it highlights the correspondence between texts (Yousef and Janicke, Survey, 2021). Ugarit provides a side-by-side view that gives immediate response to the aligned tokens for both editing and viewing (Yousef 2020). Ugarit also generates bar charts for statistics of alignment pairs, useful for an overview on the granularity of alignment (Fig. 1). For translation studies, the most useful may be the chart-view of aligned pairs, which provides a list to display how different translations render the same token from the original language, or the lack of direct correspondence i.e. omitted by the translation. The presence of alignment between only translations but without correspondence to the original text can be especially telling of the translation process, which will be discussed later.

Fig. 1: Table-view of translation pairs on Ugarit, downloadable data.

Doing text alignment provides an opportunity to review metrical composition otherwise hard to detect in an unknown language. The passage in question contains a verse section in Sotadean meter, which is an important formal observation as it adds a new layer to the complexity of the Satyricon’s embedded literary tradition. The Sotadean metre inverts the Homeric dactylic hexameters and is initially used for obscene or parodic verses, which gives more literary nuance to Sat. 132.2 (my alignment), as it gives an epic narrative to Encolpius’ failed castration in shame (Sapsford 2022, 3; 157). Both translations fail to capture the exact meter; but ancient metre getting lost in translation has been a long-term issue, that the compromise is taken as unavoidable (Crane et al. 2023, Beyond Translation, 170). While secondary literature may highlight the lost metre, in-text annotation that highlights syllables of different length works more efficiently, such as metric annotation for Homeric texts by David Chamberlain, though the annotated corpus is limitedly available. Translation alignment shows the correspondence intertextually between pairs, but this does not stand in for metrical data, though the visual correspondence to some extent highlights linguistic aspects such as the position of verbs which may have significance in the reader’s understanding of poetry (Fig. 2). However, despite its verse form, translation A has a much lower alignment rate than B. This may not be noticed without alignment because most imageries are kept intact in A (where all three texts are aligned) and provides the impression that both translations are as loyal to the original. But A invents and omits parts of the texts that undermines the textual integrity which the original Latin depends on to form intertextuality. By processing which Latin texts are aligned, the alignment highlights clearly which parts are translation’s invention or omission.

How does each translation process the question of the Satyricon’s form is a question even more relevant with the Satyricon’s non-verse sections. Translation alignment shows that, in the beginning of Sat. 132.1, translation B follows the parallelism in Latin, beginning with iam (now); but translation A abandons this structure and merges the actions into one continuous train of actions with ‘our’. This alteration may seem minor but though A keeps to the parallelism that begins each section of a sentence with the same word, the word is unaligned in Latin, suggesting that A put emphasis on contemporary readability that treats the Satyricon as a novel for entertainment. In contrast, Translation B is highlighted for a word-to-word approach as shown by the unaligned parts which are largely adverbs and subjects, already suggested or inflected in Latin.

However, unaligned texts, even adverbs and subjects, should not be all dismissed as simply semantic differences. Looking at the unaligned Latin words—Sat. 132.2: illa, singular ‘she’, is used to refer to Encolpius’ penis. The feminine pronoun holds significance in Roman semantic context but is missed in both translations. In Sat. 132.2, Translation A omits fugi (I fled), and its implication on the mutual cowardice in Encolpius and his genitals, which tweaks the dynamic of the mock-epic castration from self-shaming to a relentless pursuit to combat his impotence.

Fig. 2: Visualisation of alignment between original verse, non-verse translation and verse translation.

The side-by-side view between translations may reveal critical differences between versions. In the beginning of 132.4 (Fig. 3), translation A and B use opposite pronouns, one female (B) one male (A) for Encolpius’ penis, and they are carried on down the verse. Similarly, the word solo in 132.4 is understood as ‘soil’ in B and ignored in A, which alters the meaning. Translation A likely makes the change for the reason already argued: to make the text comprehensible in English, not so much to reflect the Latin semantics.

Translation B’s approach may be more efficient for non-Latinist classicists to work literary criticism with. However, words that rely on Latin’s inflected structure or carry meanings embedded in Latin literary semantics are still compromised in translation for the sake of coherence within English syntax and usage. The noun venereum has been translated to ‘make love’ (B) and ‘love’ (A), which are both semantically coherent but omit the evocation to Venus in Latin. In Sat. 132.3: furciferae is an altered form of the invective furcifer that’s commonly used on slaves, again, actively aligning the text may direct the reader to the untranslated servitude imposed with invective. In Sat. 132.1, catomizari is an unusual conjugation for catomidio that cannot be found in Latin online dictionaries. Both translations give it the exact wording ‘hoisted and flogged’. During the alignment, the curious absence of the word from dictionaries may direct the reader to secondary literature, which would lead the discourse on multiple versions of Latin texts, raising awareness to annotations for Latin that may otherwise not be as significant to a non-Latinist’s knowledge.

Fig. 3: The presence and absence of the feminine pronoun, highlighted.

By practicing active translation alignment, one is naturally drawn to the challenges. As demonstrated above, unaligned text or contradicting meanings between translations often flag the ‘untranslatability’ and directs to more in-depth understandings. Between alignment of translations, we may also learn about which ‘original’ text each translator works from. The end of 132.3, according to Loeb’s Latin is ‘rogo te, mihi apodixin « non » defunctoriam redde’. Alignment indicates a level of correspondence with both translations but conveying very different meanings: A asking for ‘sign…however faint’ and B for ‘serious proof’. It suggests that the two translations may be using different Latin texts, especially with the ambiguous addition of ‘« non »’. As made identifiable by the list of translation pairs, Sat. 132.4 has alignment occurring only between translations: ‘words’ is included for translations but not found in Latin. It may suggest cross-referencing between the later to the earlier version, or the same translation convention has been used.

My project conducts a practical manual translation alignment with an ancient text and two temporally distant translations, in evaluation of how active alignment can assist with what would have been the persistent issue of untranslatability for people without knowledge of the ancient language(s). While demonstrating that the 2020 translations have an emphasis on conveying meanings as exact as possible, as coherent to the scholarly recognition of the Satyricon, the 1922 translation prioritises the Satyricon as both entertaining and equivalent to a modern novel. The experience of manually aligning the texts is proven useful for a non-linguist to give a more semantically inherent—a sense of lingual ‘nativeness’—analysis to their unknown languages. From the interactive alignment practice aided by immediate visual response, both aligned and unaligned texts are useful for discussions. The alignment rate tells more nuance than how literate the translation is to the original. By doing active alignment, the user can efficiently spot addition and omission in translations, both in form and content, including specific tokens to detect translational aspect such as censorship. Furthermore, vast difference between translations could imply different versions of ‘original’ texts being used, directing attention to the manuscripts where the Latin texts are sourced, an aspect not often addressed by publications of translations. Including alignment as part of the literary analysis can redirect non-linguists to research for secondary annotation and literature, as alignment flags up specification that helps to direct their search in an otherwise convoluted. While doing manual alignment gives the advantage of promoting a highly focused interaction with the text, its time-consuming nature deems it more suitable for shorter passages. For longer or entire-work alignment, there are pre-aligned versions for canons such as Homer, and auto-alignment tools, such as the Ugarit auto-aligner to conduct the work for the others. As important and information-rich as the overview on alignment data is, having first-person experience with manual alignment is both a methodology to evaluate translation and a gateway to further research that gives great nuance to the data.

Posted in SunoikisisDC, Tools | Tagged report, text alignment, translation alignment, Ugarit | Leave a comment

Report on the Erasmus+ Blended Intensive Program (BIP) “Intensive ENCODE: Digital Competences in Ancient Writing Cultures”

Posted on 2025-09-23 by Martina Filosa

Introduction

Between April and May 2025, the University of Parma hosted the Erasmus+ funded Blended Intensive Programme (BIP) “Intensive ENCODE: Digital Competences in Ancient Writing Cultures.” This initiative is a spin-off of the ENCODE project (2020–2023), an international collaboration that aims to bridge the gap between traditional humanities education and the digital skills increasingly essential for the study of ancient written cultures.

Structured in two phases, the programme combined online learning (1 April–12 May) with an in-person week in Parma (18–24 May). It brought together around 70 Bachelor’s and Master’s students from across Europe, representing partner institutions in Germany (Cologne, Leipzig, Würzburg), Norway (Oslo), Italy (Bologna), Greece (Komotini), Spain (Madrid), Lithuania (Vilnius), and Bulgaria (Sofia).

What united this diverse group was a shared curiosity: how can digital tools reshape the way we explore the texts, scripts, and writing systems of antiquity?

In the blog post that follows, Elena Di Giorgio, Giulia Contesini, Alberto Negri, Violina Hristova, Stephanie Daneva, Nicoletta Nannini, Stephania Daviti, and Athena Mega, participants in the BIP, share their thoughts on some of the key themes that emerged from the guest lectures and collaborative activities.

Participants of the Erasmus+ BIP "Intensive Encode" on a stair

Fig. 1: Students and teachers of the Erasmus+ Blended Intensive Programme (BIP) “Intensive ENCODE: Digital Competences in Ancient Writing Cultures”

Relationship with writing

Nowadays, Artificial Intelligence is revealing itself as an essential tool for many tasks pertaining to any kind of field. When it comes to historical studies, AI might be trained for the purpose of automatic recognition of texts.

Pursuing such a goal, Isabelle Marthot-Santaniello applied Deep Learning-based methodologies to papyri. In the D-scribes project she worked towards identifying all the scribes responsible for the notarial documents of the archive of Dioscoros of Aphrodito. To do so, first and foremost she established a ground truth, consisting of a preliminary dataset of images representing the known handwriting of each scribe. These samples were narrowed down to καὶ-s along with some single letters (ε, κ, μ, ω), which display marked palaeographic features and show few, if any, variations in their ductus. With these specimina a confusion matrix was produced, a table that is used to define the performance of a classification algorithm. Currently, this modus operandi works well with inter-writer discrimination, but struggles with intra-writer variations. The same principle was applied to dating papyri.

The idea of automatic letter recognition was applied also in Egyptology. The Demotic Palaeographical Database Project led by Franziska Naether allows users to draw Demotic signs, so that the AI tool can attempt to identify them, even suggesting possible matches. In terms of Unicode, hieroglyphs are not yet fully represented; moreover thousands of signs, such as those found in Demotic and Hieratic scripts, still need to be accounted for. A similar challenge exists in Mycenaean studies. As a result, scholars in both fields continue to rely primarily on transliterations.

Moreover, Jérôme Mairat illustrated how helpful AI proved itelf in modeling digital editions in Numismatics (RPC online). He used AI to generate interpretations, and even translations, of Roman coin legendae, brief but dense inscriptions containing several abbreviations. Remarkably, a fine-tuned version of OpenAI’s GPT-4o-mini proved capable of recognising both Greek and Latin characters, delivering promising results with an estimated 95% accuracy rate. Of course, occasional errors and the well-known phenomenon of AI “hallucinations” still occurred. To refine the output, he employed APIs to automate the creation of both a textual edition (following Leiden conventions) and a basic EpiDoc XML file. For these more mechanical tasks, he favours using a dedicated web app, which is both more cost-effective and more reliable, unlike OpenAI, it performs the job without error.

Non-textual aspects of the writing support

The material characteristics of ancient artifacts bearing text on their surfaces must be duly taken into consideration within historical disciplines concerned with the study and interpretation of the text itself.

Papyrology, for example, is a discipline marked by a high degree of fragmentation of writing supports, and for this reason it requires constant comparison between artifacts with similar content, in order to reconstruct missing text in any existing gaps. A similar approach, though less pronounced, also applies to epigraphy in the assessment of missing or hard-to-read portions of stone.

These specific aspects of studying ancient sources are by no means marginal in DH; on the contrary, the digitization of such sources calls for methodological reflection. A useful point of reference is the Digital Marmor Parium project, led by Monica Berti. In working with this epigraphic document, particular attention is paid to the editorial layer during the encoding process, especially in areas where the text is fragmentary. When reconstructing missing portions, it is crucial not only to propose plausible readings but also to clearly indicate the editorial choices made and the degree of certainty involved.

Among the non-textual data of the writing support, iconography also occupies a place of common interest. A relevant example of the attention dedicated to this field of studies is the Orasis project, one of the digital epigraphy projects developed in Bulgaria which were presented by Elina Boeva. Orasis is a digital platform for visualizing and presenting inscriptions from monuments of Christian art from the Byzantine and post-Byzantine period on the territory of Bulgaria and other Balkan countries. In this useful resource, it is possible to find a detailed description of the location of the image in the context of the iconographic program of the church.

Two more examples of digital endeavors in the humanities are worthy of mention. The first is the database MetrICa (Metrical Inscriptions of Campania) carried on by Pietro Liuzzo. MetrICa aims at investigating the relationship between the text and its material supports, graphic rendering, paleography, and the original archaeological context, when it can be reconstructed, emphasizing the importance of reuse of open data and collaboration.

Digital Scholarly Editions

In recent years, numerous projects in the field of Digital Humanities have emerged and evolved, each dedicated to different types of artefacts from antiquity. The diversity of research objects, project goals, and established scholarly traditions requires a range of methodological approaches, sharing a common objective: to produce high-quality, scholarly, and accessible digital publications, following the FAIR principles.

The already mentioned Digital Marmor Parium project, for example, uses annotation and named entity extraction to build a database designed not only for specialists in linguistics and literature, but also as a resource that can be integrated into broader research initiatives. The process of extracting this type of data from epigraphic monuments and papyri presents several challenges: not all resources and authority lists are openly accessible for instance, Trismegistos, and others suffer from structural issues.

One particularly ambitious project is The Digital Rosetta Stone by Franziska Naether, Monica Berti, and their team, which combines 3D models of the monument, high-resolution imagery, visual alignment of the hieroglyphic, Demotic, and Ancient Greek texts, and tools such as Treebanking, semantic annotation and text alignment (Ugarit).

Similarly, the Damos project by Federico Aurora, a database of Mycenaean inscriptions, demands a tailored approach to text encoding due to the unique features of Mycenaean Greek. When encoding these inscriptions using EpiDoc, numerous difficulties and inconsistencies arise in comparison to encoding inscriptions in Classical Greek or Latin. A notable distinction lies in the requirement to annotate each word explicitly in Mycenaean texts, since some are represented by logograms a necessity not present when working with texts in Classical Greek or Latin.

In Bulgaria, key examples include the Tituli (Latin inscriptions), Telamon (Greek inscriptions), and Orasis (Byzantine texts found in churches), presented by Elina Boeva. The projects focus on producing digital publication of the inscriptions, using an EpiDoc template, tailored specifically for the needs of each project.

The model of digital publishing increases accessibility, allowing broader public and academic engagement without requiring physical access or paid subscriptions. However, it also comes with challenges, including the ongoing need for institutional funding to support hosting, maintenance, and database updates.

Content and meaning

Using digital tools for the study of Greek and Latin Epigraphy and Papyrology aims to offer inspiring content, fostering awareness of the importance of digital competencies, training in the field of research, as well as study of ancient writing cultures. The analysis brings out a range of linguistic and semantic aspects that were previously difficult or time-consuming to explore and promotes the digital transformation of cultural heritage by bridging the disciplinary gaps.

Digital technologies have significantly advanced paleography and script recognition. Computer vision techniques facilitate the identification of characters in damaged manuscripts, while machine learning models enable precise dating of handwriting styles such as uncial and cursive scripts. Morphological analysis benefits from automatic lemmatization and tagging, particularly in inflected languages like Ancient Greek and Latin. Additionally, digital syntax parsing improves reconstruction of incomplete structures and detection of formulaic expressions in legal and religious texts. These methods also support the study of linguistic variation, including dialectal differences, orthographic shifts, and code-switching phenomena.

Moreover, digital tools play a crucial role in exploring semantic aspects of ancient texts. Named Entity Recognition (NER) automatically identifies personal names, places and institutions, facilitating the construction of prosopographical databases by linking individuals across documents. Topic modeling and thematic clustering enable the detection of recurring themes, such as taxation, marriage, or military service, across large corpora, supporting semantic classification and contextual analysis. Tools for intertextual and citation analysis uncover textual reuse, quotations, and allusions to earlier literary, legal, or religious sources, shedding light on intellectual networks and the transmission of knowledge over time.

Digital projects analyzing ancient texts increasingly employ computational methods to investigate linguistic and semantic features. Named Entity Recognition (NER), used by Trismegistos, Papyri.info and the Digital Marmor Parium project identifies and standardizes personal and place names, aiding prosopographical and geospatial analysis. Morphosyntactic annotation, as in the Perseus Digital Library, addresses the complexity of inflected languages like Ancient Greek and Latin. Semantic disambiguation techniques, such as word embeddings and topic modeling (e.g., Tesserae), resolve lexical ambiguity. Thematic clustering, seen in the Digital Corpus of Literary Papyri, supports text classification. Projects like Pelagios emphasize intertextuality and citation networks. Multilingual corpora (e.g., Greek-Coptic) and infrastructures like CLARIN, DARIAH-EU, and ENCODE further enhance scholarly research.

Linked Open Data and Semantic Web technologies are crucial for digital analysis of ancient texts. EpiDoc, based on TEI-XML, standardizes encoding of inscriptions and papyri with rich metadata. Projects like Trismegistos use RDF to link entities via ontologies such as Pleiades and SNAP:DRGN, promoting integration. The Pelagios Network connects texts to geographic data for spatial analysis. Tools like Papyri.info Editor and Recogito support collaborative annotation, while databases such as EDH and EAGLE offer structured, API-accessible resources enhancing interoperability and epigraphic research.

Variety of languages, chronology, and civilizations

Studies of ancient civilizations must contend with a history made up of places, chronologies, and, above all, extremely diverse languages. When these fields of study intersect and engage with the domain of digital humanities, the result is the need for digital tools with specific features: they need to be capable of conveying the diversity of content, depending on the contexts under examination, in a way that is as simple as possible for developers to use and as accessible as possible for the audiences who rely on them.

One of the most significant aspects in the encoding of ancient texts lies not only in the goal of making the texts increasingly accessible, but also in connecting features that may be shared across different languages and civilizations, as through the analysis and study of linguistic aspects.

For instance, it is precisely from this perspective that was created the WordNet project, a large lexical database or “electronic dictionary”, developed at Princeton University for Modern English. The aim of WordNet is to collect and interlink not just words based on their meanings, but specific senses of words. WordNet labels the semantic relations among words and provides information about two fundamental properties of the lexicon: synonymy and polysemy. In the study of ancient languages, it proves especially useful to link WordNets with other textual and lexical resources and it reveals connections between the semantics of lexical items and their syntactic context. Especially in philology, this connection may help researchers filling gaps in the written records, as entire sets of near-synonymic and semantically related words will be easily made available.

Furthermore, the project faces a pivotal challenge: whether to adopt English synsets as a foundational structure an approach that risks significant inaccuracies or to construct an entirely new synset framework, which would require extensive linguistic and technical effort. Additionally, ancient languages contain conceptual distinctions that are not lexicalized in English, e.g. avunculus and patruus refer to maternal and paternal uncles – distinctions not explicitly encoded in English.

Posted in EpiDoc, Events, report, Teaching | Tagged digital classics, epidoc, epigraphy, linked open data, papyrology, teaching | Leave a comment

University of London Digital Humanities Placement report

Posted on 2025-06-04 by Gabriel Bodard

By Shibo Wu and Matthew O’Connor

This spring we, two students of the Digital Humanities MSc at UCL, undertook a three-week placement at Senate House Library and the Makerspace in the University of London, digitising artefacts from the Harry Price Archive (part of the Harry Price Collection) in Senate House Library under the supervision of Gabriel Bodard and Tansy Barton. This placement was part of an ongoing initiative to enhance the accessibility and preservation of rare and fragile items in University of London archives.This marked the continuation of the digitisation initiative that began with the Ehrenberg Collection at the Institute of Classical Studies, introducing the first digitised artefacts from the Harry Price Collection.

The placement commenced with two days of inductions to both the Ehrenberg Collection and the Harry Price Collection, familiarising us with their historical significance, cataloguing practices, and handling procedures. The Harry Price Collection was established by the psychical researcher Harry Price (1891-1948), a contemporary of Harry Houdini, containing articles and artefacts related to magic, spiritualism and psychical research. Donated to the Senate House Library in 1936, and bequeathed to the University of London in 1948, the collection includes various items from Price’s own investigations into occult and/or spiritual phenomena, as well as random objects relating to his antiquarian and other interests.

We conducted several photography sessions to capture high-quality images of select artefacts in the Harry Price Collection. These items were placed on a rotating turntable, photographed with a Nikon D850 digital SLR camera, and imported into the Agisoft Metashape Pro software for processing. We were able to produce 3D models for our artefacts using this software, which were then uploaded onto the sharing platform Sketchfab.

Some of these items, particularly those with metallic surfaces, proved challenging to model due to inconsistent reflections which disrupted the software’s alignment for images, even with adjusted lighting, exposure and the use of a light tent. Other items were deliberately symmetrical or featureless because they were designed as tools for magic tricks, therefore offering few reference points for marker-based alignment when merging chunks together. We learnt from these less successful attempts, allowing us to mitigate or avoid these issues when dealing with subsequent artefacts.

Silver Ingot with Latin Engraving by Harry Price Archive on Sketchfab

This early troubleshooting allowed us to develop careful and creative solutions for the artefacts we uploaded to Sketchfab; namely the Silver Ingot, the Metal and Glass Flashlight, the Lead Imitation Figurine and the Rapping Hand. For the Silver Ingot, the model was divided into front and underside chunks, manually merging them using precise marker placement in order to preserve the engravings on the front chunk (above) (direct link). We adapted this method for the Metal and Glass Flashlight, which featured a fractured glass side that distorted the model’s geometry (below) (direct link). To resolve this, we created separate low-resolution chunks for each half of the object, then constructed a clean base geometry onto which smaller high-quality chunks could be accurately aligned.

Metal and Glass Flashlight by Harry Price Archive on Sketchfab

A micro lens was used to capture the finer surface details of both the Lead Imitation Figurine and the Rapping Hand, allowing for improved clarity on intricate features such as the figurine’s face and the lace of the hand’s velvet cuff. The micro lens also picked up background particles, however, requiring significant manual cleaning in post-processing. In particular, the Rapping Hand required extensive manual and automated processing, including masking, cleaning, and smoothing to address its complex geometry. Due to these processes, and upload constraints, the final model had to be decimated and compressed, which resulted in a noticeable loss of accuracy. By contrast, the Lead Imitation Figurine (not from the Harry Price Archive) benefited considerably from the micro lens, with its surface definition resulting in a highly-detailed 3D model upon upload (below) (direct link).

Lead Imitation Figurine by Harry Price Archive on Sketchfab

Gabriel provided us with a detailed workflow document, updated by placement students over the past several years, to which we also contributed our notes, documentation and paradata. This served two crucial purposes: it ensured that mistakes were recorded systematically, highlighting where the process had faltered, and acted as a reference for future digitisation efforts. This workflow offered guidance from previous projects when we met with significant issues, and allowed us to present our own insights which might benefit future teams. By documenting each step in the process, including our failed attempts, we contributed to a growing repository of practical knowledge.

We are immensely proud to have contributed the inaugural items to this new digitisation collection, setting the foundation for expanded digital access to the Harry Price Archive. We hope that these initial models will serve as a stepping-stone for future viewers and contributors to engage with these artefacts remotely, preserving their significance for decades to come. We are deeply grateful to Gabriel Bodard, the Institute of Classical Studies, Tansy Barton, Salvador Alcantara Pelaez and Senate House Library for providing us this incredible opportunity, and we look forward to seeing this digital archive grow in the years ahead.

The models uploaded to the Harry Price Collection on Sketchfab can be found here.

Posted in Teaching | Tagged 3d imaging, digital humanities MSc, occult, photogrammetry, student placement | Leave a comment

Digital Classicist London 2025 programme

Posted on 2025-05-01 by Gabriel Bodard

This year’s Digital Classicist London summer seminar series, hosted by the Institute of Classical Studies, University of London, invited proposals on any aspect of the ancient or pre-colonial worlds, including archaeology, cultural heritage, history, language, literature and their reception, that address innovative digital approaches to research, teaching, dissemination or engagement. We particularly welcomed proposals that integrate research, archival and scientific data across disciplinary or other boundaries.

The organisers (Gabriel Bodard and Elizabeth Koch-Kölük at the ICS, Stephen Kay at the British School in Rome, and Katharine Shields at King’s College London) are delighted to share the programme of seminars, all of which will be streamed live as well as delivered in person in London. The rich variety of presenters, methodologies and subject areas represented in this programme is especially pleasing to us, and helps to showcase both the depth and breadth of this thriving discipline.

All seminars are held at 17:00 BST (UTC+1) on Fridays, live on Youtube and in person in Senate House MakerSpace, room 265, University of London.

Booking recommended for in-person attendance. Not required for online attendance (live or any time after) at Youtube. Use links below.

Friday June 6: Matteo Romanello (University of Zurich) & Charles Pletcher (Tufts University), Introducing Kōdōn, a Minimal Computing Library for Publishing Digital Commentaries (youtube) (register)
Friday June 20: Valentina Lunardi (University of California, Los Angeles) & Barbara McGillivray (King’s College London), Static and contextual embeddings for tracing semantic change: the case of Christian Latin (youtube) (register)
Friday July 11: Thibault Clérice (Inria, Paris), Distributed Text Services for Digital Classics (youtube) (register)
Friday July 18: Ester Salgarella (University of Aarhus), Sort It, See It, Say It. Digital strategies to revive a Bronze Age Aegean Script (youtube) (register)
Friday July 25: Chiara Senatore (La Sapienza, Roma), Digital editions of classical texts for GLAMs (youtube) (register)

We look forward to seeing you there!

Posted in seminar | Leave a comment

Material Digital Humanities seminar 2025

Posted on 2025-03-07 by Gabriel Bodard

The Material Digital Humanities seminar is organised by Gabriel Bodard (Digital Humanities, University of London, UK) and Chiara Palladino (Department of Classics, Furman University, USA) in 2025. This series presents a range of discussions around materiality and the research possibilities offered by digital methods and approaches. Beyond just the value of digitization and computational research to the study of material culture, we are especially interested in theoretical and digital approaches to the question of materiality itself. We do not restrict ourselves to any period of history or academic discipline, but want to encourage interdisciplinarity and collaborative work, and the valuable exchange of ideas enabled by cross-pollination of languages, areas of history, geography and cultures.

For this year’s seminar, the organisers have arranged a conversation between several of the contributors to a volume that appeared a year and a half ago, on the relationships between materiality and the digital in cultural heritage, and other scholars working in related areas.

Palladino C. & Bodard G. (eds.) 2023. Can’t Touch This: Digital Approaches to Materiality in Cultural Heritage. London: Ubiquity Press. DOI: https://doi.org/10.5334/bcv

As with previous years, this series represents a wider conversation around issues relating to material culture, heritage and materiality and the ways in which digital methods and the discipline of digital humanities engages with them. Even traditional areas of DH focus including text, images and (born-)digital data suffer from ignoring their material and other contexts, be they spatial, chronological, scholarly engagement, ethical and historical settings.

Please do join us and contribute to these discussions!

Wednesday March 5, 2025. Rebecca Kahn (University of Vienna), Skulls, skin and names: The ethics of managing heritage collections data online (Youtube)
Monday March 10, 2025. Dipali Mathur (Ulster University, Belfast), The Limits to Green Growth: Rematerialising UK’s Pathway to Net Zero
Monday March 24, 2025. Chijioke Okorie (University of Pretoria), Digital treatment of African cultural heritage: Shifting landmarks and implications for copyright exceptions for archives (Youtube)
~~Friday May 2, 2025~~. Martina Filosa (Universität Köln), Usama Gad (Ain Shams, Cairo), Gabriel Bodard (University of London), Making the Implicit Explicit: Digital editions of ancient text-bearing objects (POSTPONED – will reschedule after summer)
Tuesday May 13, 2025. Rita Lucarelli (Berkeley), Ancient Egypt Reimagined: Teaching about the Past Through Digital Humanities (Registration)
Tuesday June 3, 2025. Francesco Bianchini (Cambridge University), Sanskrit inscriptions as tangible artefacts: exploring digital tools (Registration)

All seminars are free and online only. Please register for the zoom link.

Full details and updates are in the SAS Digital Humanities calendar.

Posted in Events, seminar | Tagged digital humanities, MaterialDH, materiality | Leave a comment

Online Tools for Handwritten Text Recognition: A Comparative Review of Transkribus and eScriptorium for Byzantine Paleography

Posted on 2025-01-21 by Martina Filosa

written by Konstantina Eleftheriadi, University of Cologne

Introduction

This article investigates online tools for Handwritten Text Recognition (HTR). Specifically, it examines and compares the functionality of the “Transkribus” and “eScriptorium” platforms. Initially, it outlines practical applications of either platform. In the case of Transkribus, a custom Deep-Learning HTR model for Byzantine Greek is developed and evaluated. In the case of eScriptorium, an established model is tested instead, for the same tradition. Through the two experiments, the efficacy of each platform is revealed. Consequently, there is an attempt to compare them briefly, before ultimately venturing to highlight prospective challenges that arise regarding the applicability of such tools for research in the field of digital paleography.

The framework for the endeavor has been provided by the “Sunoikisis Summer 2024: Digital Classics and Byzantine Studies” Consortium, convened by Martina Filosa, Monica Berti, and Gabriel Bodard, for the Academic Summer Semester 2024. Specifically, the fourth session from the aforementioned, on “HTR and OCR from papyrus to codex” and the four proposed exercises.

Online Tools for Handwritten Text Recognition

Manuscripts constitute vital resources for the formation of historical narratives, insofar as they are direct carriers of information concerning past sociocultural circumstances (Saraswati 2013). The study of the material may often be impeded by practical conditions, chief amongst them being the fragility of the objects and their dispersion among different collections and locations. Digitization efforts have significantly increased accessibility to documents, motivating their comprehensive study remotely (Marthot-Santaniello 2021). The development of highly-specialized tools that enable innovative approaches to the new, digital material in academic contexts, has also helped invigorate interest in their study (Jang 2020). “Digital Paleography” has emerged as prolific scientific direction in its own right, garnering due attention from experts.

Transkribus: Creating a custom HTR-Model

The second and third exercises provided instructions for the development and evaluation of Deep-Learning models for Handwritten Text Recognition in Byzantine paleography. Excerpts from manuscripts of the Bodleian Library Collection were supplied in the form of JPEG images, along with transcriptions of their contents, to be used during the training and validation stages. The process was conducted through the “Transkribus” platform, co-developed by the University of Innsbruck (A Short History, 2023). As per the directions of the second exercise, seven pages in total were used in the endeavor. The pages selected originated from the MS. Barocci 102, a Commentary on Isaiah by Basil the Great, compiled around the early 12^th c. The model was trained on six pages, then validated and tested on two more.

In the first stage, the model used the training pages and their ground-truth transcriptions as an example, to begin learning to recognize character-sequences in the images. Afterwards, it evaluated and fine-tuned itself by testing its accuracy against the validation-dataset, for which it also had access to transcriptions. Finally, it could attempt to transcribe the text on the test-page, for which it did not have a transcription. The entire process was conducted automatically through the platform, with no further intervention by the user. A preliminary examination of the results on the test-dataset could immediately point to disparities from the text (Fig. 1). Especially concerning was the misinterpretation of word-clusters in areas where the text was densely written, meaning with insufficient spacing between glyphs.

Fig. 1 – Transkribus: Page transcribed utilizing the created model.

The experiment helped further highlight the difficulties cited in literature, concerning the development of functioning models for HTR implementations (Pavlopoulos et al. 2023A). Peculiarities in the handwritten text impeded the recognition and transcription processes, leading to an array of mistakes. Deviations in positioning posed a particular obstacle, as confounded spacing and base-line gravity led to frequent misidentifications and incorrect collation of words. Ligatures and abbreviations were equally challenging, as the model could not separate the individual letters that formed them. As proven by the comparison of common-word instances, stylistic similarities among glyphs also served to confuse the algorithm, as υ (ypsilon) and ν (ny) were often interchangeable. The issues presented could possibly explain the disparity in detected characters, as well as the significant Character Error Rate Percentage (%CER) calculated (Fig. 2, 43.99%).

Fig. 2 – Transkribus: Presentation of the created model.

eScriptorium: Using a pre-trained HTR-Model

The fourth and final exercise provided a pre-trained model for layout recognition and automatic transcription of Byzantine Greek Handwritten Manuscripts of the 10^th c. A concise workflow was outlined in the instructions of the exercise, for the employment of the model in the framework provided by the “eScriptorium” platform, developed by the École Pratique des Hautes Études (Stokes et al. 2021). Though the former may be downloaded locally on Linux or Mac computers, an account on the server of the CREMMA Project was also made available, in order to access the application. Thus, the inability to access the application on a Windows operating-system, could be overcome. The latter option was favored, as it allowed for the processing and storing of information on a cloud server, minimizing the demand on the student’s operating system. For the purposes of the exercise, three pages of the Palatinus graecus 23 from the Bibliotheca Palatina were selected to be examined. An examination of the produced text showed promising results, in terms of fidelity to the original. Overall, the transcription was satisfactory and would require significantly less post-correction, compared to the ones produced previously (Fig. 3).

Figure 3 eScriptorium: Visual comparison of recognized transcription on p. 49 of the Codex Palatinus graecus 23. On the top, the first 8 lines on the page, as segmented and transcribed automatically utilizing the provided model, on eScriptorium. On the bottom, the page as segmented on the digital Bibliotheca Palatina.

Fig. 3 – eScriptorium: Visual comparison of recognized transcription on p. 49 of the Codex Palatinus graecus 23. On the top, the first 8 lines on the page, as segmented and transcribed automatically utilizing the provided model, on eScriptorium. On the bottom, the page as segmented on the digital Bibliotheca Palatina.

Platform Comparison

Transkribus and eScriptorium may be thought of as ultimately accomplishing similar results, as two examples of highly specialized platforms. Both were developed in an academic context and have been regularly employed for robust scientific work. Both were judged as having a user-friendly interface that enabled engagement with their tools with relative ease. The steps that had to be followed for each procedure were clearly communicated through labeled buttons and dedicated panes. Transkribus especially offered extensive instructions in the “Help Page”, which further enhanced the experience. Waiting times were similar, in the range of 15-25 seconds for segmenting, and for transcribing documents.

A major disparity between the applications would relate to actual accessibility. Transkribus is a licensed platform that requires a paid subscription to be utilized to its full potential. Currently, it only operates through a provided server, wherein operations are confined to what is allowed as part of the running payment plan. The API behind the platform is only accessible to collaborating organizations. Conversely, eScriptorium is open-source and may be freely accessed on Linux and Mac operating systems, or through a cloud server. Its full documentation is available online, along with its API. Community support is highly encouraged, in order to consolidate the platform. A final difference between the two is that eScriptorium also allows the export of models trained on the platform, ensuring the shareability of its products. Thus, the results of a scientific work may be replicable in new contexts.

Discussion and Challenges

The digital publication of produced data is steadily becoming a sine qua non for the investigated disciplines (Galleron – Idmhand 2020). Advancements in relevant software have permitted complex interactions with the material in varied environments. Cloud-based operations are becoming more of a norm, limiting the burden on private operating machines. The compounded engagement of experts with the new methodologies has opened up a new epistemological horizon for the disciplines, in the era of “Digital Humanities”. By examining the prospects and limitations of the tools presented, major challenges in the realm of Digital Paleography may also begin to be delineated.

The threat of data obsolescence remains a significant concern for pertinent implementations. Projects that are no longer retrievable or emulable are often encountered, as no measures were taken to ensure their permanence in digital form (Strange et al. 2023). Informed data management strategies should be deliberated over, to ensure that the models developed and the archives they are included in are not only functional in the present, but also replicable in the future. Interoperability may be set as an indispensable condition in publishing related work for posterity, to motivate its continuous refinement through the input of the scientific community. Consequently, “digitization-bias” should be taken into account when interacting with relevant tools. Datasets may be limited to what has been digitized, or simply preserved in accessible digital form, thus affecting the analytical scope of scientific research that heavily relies upon digital databases.

Additionally, the employment of digital tools such as the ones surveyed for research and analysis en masse, could ultimately result in the “black boxing” of scientific knowledge. In the field of Digital Humanities, the former term is understood as a process in which the over-dependance on software to perform complex queries contributes to limiting the user’s understanding of essential background operations (Stokes 2009). Procedures in digital environments often require a sequence of operations in advanced mathematics and statistics to be carried out, as vital components that make up the deep learning algorithms enlisted. Simultaneously, an efficient tool that produces a desired result, may not motivate scrutiny from the researcher, as it serves its purpose successfully without intervention. When relying primarily on applications that simply accept input and produce an output with little clarification, those same procedures become obscured to the user, who cannot effectively re-apply them without fear of loss of scientific integrity.

Regarding the problems presented, a viable solution has emerged in the form of extensive collaborations. When dealing with such heterogenous structures, it would be expected to face an array of complications, which can be alleviated through interdisciplinary cooperation. Indeed, the implication of experts from varied backgrounds has been common-practice in related projects. Computer scientists and data analysts have often aided the work of philologists. In a reciprocal manner, philologists have motivated pioneering approaches in the aforementioned areas. Both sides have sought to expand their knowledge and toolshed as well, by specializing in digital applications for the humanities (Galleron – Idmhand 2020).

Shareability of data and analytical processes has become another central objective in modern academia. Transparency concerning tools utilized, problems encountered, and error-correcting methodology favored, additionally strengthens relevant efforts (Joyeux-Prunel 2024). By adhering to protocols that promote open-access and collaborative practices, a fertile environment for the proliferation of the disciplines may be formulated presently, and for the future. As tools become more powerful, and experts gain more knowledge through the exchange of ideas, new directions for even more creative analyses may surface. Ensuring the explainability and reproducibility of published methods motivates participatory values among the scientific community, which aim at refining the introduced epistemology. The consistent reference to the FAIR Principles (Fair Principles, 2022) for the publication of digital data, serves to exemplify that point.

This scientific paper was supported by the Onassis Foundation – Scholarship ID: F ΖT 020-1/2023-2024.

Posted in SunoikisisDC, Tools | Tagged Byzantine Studies, digital classics, Digital paleography, eScriptorium, HTR, manuscripts, palaeography, SunoikisisDC, Transkribus | Leave a comment

Digitizing Antiquities: A 3D Photogrammetry Placement at the Institute of Classical Studies

Settling into the MakerSpace Workflow

Week One: The Ceramic Sherds and the Sponge Solution

See the 3D models in Sketchfab:

Challenges and Problem Solving

Week Two: Tackling the Stone Statuette

Reflections

To cite this article:

Linked Pasts 11 Activity – Promoting Collaboration, Linking Communities

Building Pelagios connections through the Digital Classicist Wiki

Digital Approaches to Sacred Space in Roman Britain

Description

Programme

Abstracts

Fragmented Data and New Networks in the Study of Roman Religion

Speakers: Alessandra Esposito (King’s College London), Francesca Mazzilli (Università degli Studi di Bari Aldo Moro)

Mapping Sacred Spaces in Roman Britain: A New Corpus on Wikidata

Speaker: Maxime Guénette (Université de Montréal)

RomAniDat and the Pandora platform: using spatial modelling software for analysing animal bone data from temple sites

Speakers: Tony King (University of Winchester), Dominika Schmidtova (Masaryk University), Ricardo Fernandes (Max Planck Institute for Geoanthropology)

Digital approaches to space and excavation at the Roman shrine landscape of Teffont, Wiltshire, UK

Speaker: David Robert (Cardiff University)

Spectres of the Past: Haunting, Uncertainty, and Space for Digital Imaginaries

Speaker: Anna Collar (University of Southampton)

Integrating landscape-scale prospection, high-definition 3D GPR and targeted excavation at a newly discovered Romano-Celtic temple in Somerset

Speakers: Jodie Lewis (University of Bradford), Michael Pisz (University of Bradford)

Identifying sacred sites in the rural landscape

Speaker: Stephen Clifton (University of Kent)

Londinium AD 215: imagining urban sacred space in Virtual Reality

Speakers: Dominic Perring (University College London), Jake Nixon (Historic VR)

Linked Pasts 11 – Epigraphic Data in Wikidata: From Ontology to Practice

Report

Day 1 Introduction, Theoretical Framework and Hands-On Practice

Theoretical Framework: Linked Open Data and Wikidata

Data Structure and Modelling

Practical Session

Day 2 Advanced Modelling, Validation, and Case Studies

Epigraphic WikiProjects and Data Models

Breakout Sessions

Outcomes and Reflections

I. Project Outline and Aim

II. Methodology

III. Exploring EFES out-of-the-box

III.1. Indexing

III.2. Facets and Searching

III.3. Date Slider

III.4. Adding Images

III.5. Multilingualism

III.6. Bibliographic Concordance

III.7. EFES-Pollentia Homepage

IV. Discussion/Conclusion

Introduction

Fig. 1: Students and teachers of the Erasmus+ Blended Intensive Programme (BIP) “Intensive ENCODE: Digital Competences in Ancient Writing Cultures”

Relationship with writing

Non-textual aspects of the writing support

Digital Scholarly Editions

Content and meaning

Variety of languages, chronology, and civilizations

Introduction

Online Tools for Handwritten Text Recognition

Transkribus: Creating a custom HTR-Model

eScriptorium: Using a pre-trained HTR-Model

Platform Comparison

Discussion and Challenges

This scientific paper was supported by the Onassis Foundation – Scholarship ID: F ΖT 020-1/2023-2024.

Hosted by the Institute of Classical Studies

In memoriam

Recent Posts

Recent Comments

Categories

Meta

Archives

Pages

Recent Posts

Recent Comments

Categories

Search

Archives

Meta

Day 1
Introduction, Theoretical Framework and Hands-On Practice

Day 2
Advanced Modelling, Validation, and Case Studies