RAC/TRAC Session 28: Theoretical Approaches to Big Data in Roman archaeology

Conference Sessions and Abstracts - Saturday 13 April 2024

28. Theoretical Approaches to Big Data in Roman archaeology

Penny Coombe – Getty Research Institute; University of Reading
Nicky Garland – Archaeology Data Service

Roman archaeology has long produced large and complex data. Creating, managing, and sharing ‘big’ datasets has been of perennial interest for many archaeologists. The potential and limits of ‘big data’ have been recently highlighted and discussed and a light shone on how digital heritage relates to recent research in Roman Archaeology (e.g. Garland TRAC webinar 2023; TRAC 2023 Digital Archaeology workshop; TRAC 2022 session 4). Technological advances and increased application of data principles in Roman archaeology provide the urgent impetus and opportunity for critical reflection on the theoretical approaches that underpin these analyses. By identifying the theoretical frameworks that drive the production, use and reuse of big data in Roman archaeology we can better understand the potentials and pitfalls of these approaches.
This session provides space for theoretical analysis. In particular, we welcome papers and discussion on, but not limited to:

• What assumptions have been made in constructing datasets and ontologies?
• Can the theoretical discussion of data and big data (e.g. Morgan 2022) elucidate more aware and humanised interpretations?
• What common principles for compiling, combining and sharing data are needed?
• How can we combat the ‘siloisation’ of data within archaeological subfields (Lawrence 2022)?

Bibliography:
Lawrence, A. 2022. Harder – Better – Faster – Stronger? Roman Archaeology and the challenge of ‘big data’. Theoretical Roman Archaeology Journal 5 (1): 1-29
Morgan, C. 2022. Current Digital Archaeology. Annual Review of Anthropology 51: 213-231 

Session schedule 

Abstracts 

 Bringing coins out of the cabinet. Nomisma.org, a controlled vocabulary and ontology for numismatic linked open data
David Wigg-Wolf – Deutsches Archäologisches Institut

As more or less standardised, mass-produced serial products, coins are ideally suited to digital applications, in particular in the context of databases, linked open data and the semantic web. Since 2011 the Nomisma.org consortium has developed a controlled vocabulary and ontology for numismatics that is now widely applied in the discipline. By employing the concepts of Nomisma.org, the American Numismatic Society alone hosts data on more than 500,000 coins from 86 international projects and institutions that are accessible via portals such as Online Coins of the Roman Empire (https://numismatics.org/ocre/) and a SPARQL endpoint. Many other projects such as Coin Hoards of the Roman Empire (https://chre.ashmus.ox.ac.uk/) implement the Nomisma.org vocabulary and so are fully integrated into the wider world of linked open data. A report by the ARIADNE EU-FP7 project mentioned Nomisma.org as a particular example of good practice from which the archaeological community can benefit (G. Geser, ARIADNE WP15 Study: Towards a Web of Archaeological Linked Open Data. Version 1.0 (Salzburg Research 2016)). This paper will present the development and architecture of Nomisma.org, as well as the philosophy behind it. Questions addressed will include how Nomisma.org can – and cannot – be a paradigm for archaeological data, as well as how it can facilitate better integration of numismatic data into archaeological research and discourse.

 All Things are Numbers? Theoretical Approaches to Non-standardized Legacy Data in Romano-British Pottery Studies
Eniko Hudak – Newcastle University

Pottery is one of the most common artefacts recovered from archaeological excavations, and the variety of attributes recorded for pottery analysis produces vast amounts of quantitative and qualitative data. Over the history of the development of Romano-British pottery studies there has been an increasing emphasis on standardising the methods of and the terminologies for consistent recording of basic information to facilitate discussions. Updates of professional guidelines, however, have no scope to cover datasets preceding their publication resulting in incompatible pottery datasets across time. This paper aims to demonstrate that we should not fear the fuzzy, incomplete, and sometimes even confusing nature of legacy datasets in Roman pottery studies. The case study presented here collated big data on the distribution of Mancetter-Hartshill mortaria in Roman Britain from a variety of secondary datasets from archaeological reports of the last 60 years. It will explore the value (and perhaps even the necessity) of reframing the quantity-driven pottery analysis mindset to be able to make the most of non-standardised legacy data from secondary datasets: an application of theoretical frameworks that rely less on quantities of finds and utilises other attributes to explore province-wide research questions. Numbers may not be everything, but FAIR data is.

 Deconstructing the narrative: the use and misuse of big data
Lev Cosijns – University of Oxford        
Georgia Gould – University College London

With the increasing availability of online archaeological materials and the allure of compiling databases to store certain information together to increase its accessibility, the use of big data in archaeology has become very attractive. For example, some of the pre-requisites of certain funding bodies is to produce a database at the end of the project. However, numerous assumptions have been made in the construction of these large databases, and significant issues have often been ignored. This presentation aims to shine a light on some key issues in the building and handling of big data in archaeology using several case studies. For instance, there are significant issues which arise when small-scale archaeological surveys are compiled into one large dataset, namely a significant issue in the lack of a common language or international framework for historic chronologies, which often translates into unclear dating of sites. This can also be seen in the big data of Norwegian Migration Period textile archaeology, such as the textiles from the graves of Snartemo, Norway, which were heavily influenced by Late Roman products and trade. Thus, this presentation will deconstruct examples of larger datasets and examine the effects of big data on archaeological analysis.

 Learning from Gendered Approaches to Big Data
Alena Wigodner – Princeton University

Studying gendered behavior in the archaeological record requires us to distill a dynamic aspect of identity into static objects. Which objects represent the behavior of women? Of men? Adding to the challenge, it is easy for archaeologists to overlook the truth as our own gender biases shape our interpretations. Roman archaeology’s embrace of big data has provided new tools for studying gendered behavior, leading to what I argue is a critical symbiosis. Large datasets require us to face our gender assumptions, moving beyond the anecdotal and into the provable. Indeed, the very act of designing and populating a database shapes our approach to the relationship between an object and the gender of its user. Concurrently, gender research can guide our approach to big data more broadly, both at a theoretical level and at a highly practical one as we grapple with applying quantitative and data-driven methods to the study of a complex, intersectional aspect of identity. I illustrate the potential of this symbiotic relationship through my own database of gendered offerings from sanctuaries across Roman Britain and Gaul; my analysis especially highlights the way embracing uncertainty in dataset design and analysis helps us capture the behavior of women and other marginalized groups. Big data extends our understanding of the gendered Roman world—and vice versa.

 25,000 scattered components. A digital corpus for Roman stone architecture in Germany
Katja Roesler – Deutsches Archäologisches Institut        
Kerstin Hoffman – Deutsches Archäologisches Institut
Berenike Rensinghoff – Disjecta Membra                  
Aline Deicke – Marburg Centre for Digital Culture and Infrastructures

Roman Antiquity and its material remnants have long been the source for corpora and large data collections conducted by research institutes, like ASR, CIL, OCK, CRFB. In 2023, the new corpora project disiecta membra commenced with a reflective and decidedly digital approach to large data collections. In the next 24 years, three institutions – University Mainz conducting archaeological research, Digital Humanities at University Marburg, and RGK conducting epistemological research – with the support of several partners are going to collect and analyse 25,000 scattered components of Roman stone architecture and c. 5,000 archaeological contexts in Germany, which are hardly known and published. A specific emphasis will be on generating and modelling data for object biographies, provenance, and epistemological issues. One of the primary objectives is to publish the data in compliance with the LOUD and FAIR principles with the help of ontologies (e. g. CIDOC-CRM) and controlled vocabularies. This paper introduces and discusses our digital and structural approach as well as the epistemological research needed for building a digital collection that compiles not only data and metadata, but gives information on data creation, i.e. past sampling strategies, categorisations and (pre-)processings, and is open to future scientific questions, digital methods and technology.

What pot is it anyway? – Theoretical challenges in the classification of ‘messy’ pottery data in the Roman Northwest
Alasdair Gilmour – University of Exeter

While Roman pottery may at first seem an ideal candidate for ‘big data’ analyses, holistic approaches that would consider coarse, regional, and less standardised ceramics on an equal footing with finewares like Samian are hampered by diverse methods of classification. This is especially problematic when attempting cross-regional comparisons, and when comparing different wares recorded using different typologies and in different languages. This paper aims to discuss these inherent challenges in the production and utilisation of pottery databases in the Roman Northwest, and proposes using shape data derived from digitised typological illustrations as a potential means to address them. Lawrence (2022) notes the need for novel classifications and analytical units to properly utilise ‘big’ datasets, and by using shape data instead of typological designations and form classifications as the base units of pottery analysis it may be possible to unearth new and interesting nuances in datasets that would otherwise be obscured. Moreover, focusing on the shape data allows us to step away from our modern classifications and their implications (do we consider a vessel as a cup or a bowl?) and foregrounds the materiality of the objects themselves, something often at risk of being lost in the digitisation of archaeological material.

 Deathscapes by numbers: Exploring applications and implications of burial plot sizes from the epigraphic record
Dragos Mitrofan – University of Exeter

Burial plot dimensions indicated by the formulae such as in fronte/in agro are not uncommon in the Roman Late Republic and early Empire (1st century BCE–3rd century CE). However, these inscriptions appear prevalent in early urbanised areas of the Late Republic, and much less common in the provinces, including Britain where none have been documented. The paper has several objectives. First, it seeks to validate some of the conclusions reached by Bennet (2020) in her unpublished PhD which focused on the ECDS database. Additionally, it aims to explore other research questions, better suited for the smaller EDH database (over 600 raw data entries), such as investigating whether a more expensive material correlates with a larger plot area. Thirdly, as both Toynbee (1996) and Bennet (2020) noted, these dimensions are occasionally associated with legalistic phrases that prohibit inheritance, further partitions, or impose fines for disturbance or removal. Furthermore, with few exceptions (field boundaries, milliaria, dedications) dimensions appear to be almost exclusively present in epitaphs, raising the question whether these inscriptions have a more significant legal role to play than previously considered. The ultimate goal of this research is to construct, publish and maintain a .csv database in accordance with FAIR principles. This database will allow for a comprehensive comparison with archaeologically recorded dimensions and will contribute to a greater understanding of burial practices and their legal implications.

Back to RAC/TRAC Conference 2024 Timetable