Introduction

The present paper addresses a recurring DH-problem: how to engage with archival and legacy (archaeological) documentation that is dispersed, unevenly recorded and described in incompatible and multilingual vocabularies. Using “deviant burials” as an example from the ERC-funded project RELIC, which focuses on rural Christianisation and state formation. While the example itself is specific, the theoretical and digital framework presented is applicable beyond funerary archaeology, offering a DH approach to small, heterogeneous datasets in which the goal is not only digitisation, but making category use, comparability, and evidential context inspectable at scale.

RELIC aims to advance historical and archaeological understanding of the involvement of the general population in Christianisation and state formation, while also reflecting on disciplinary epistemic conditions in the digital age. In this contribution, we present how the data infrastructure developed for the project enables contextual comparison of controversial small datasets. Deviant burials are an excellent example of that, because “deviant” is not a stable empirical class but an interpretive label whose meaning depends on comparison with its non-deviant environment. We argue that DH infrastructures can enable such comparisons by harmonising descriptors, keeping provenance visible, and supporting cross-site exploration through interface-driven analysis.

​

Materials and Methods

The project draws on a combination of published and archival archaeological data, historical documents, and spatial datasets. Our focus is on sites that include early rural churches, field cemeteries, transitional burial grounds, and ecclesiastical and secular power centres and their settlements, i.e. the social and political context of religious change between the 10th and 12th centuries. Many of these datasets are dispersed across national and disciplinary boundaries in East-Central Europe and are difficult to compare directly due to uneven recording practices and often incompatible, multilingual descriptive vocabularies.

For the project, we employ an interdisciplinary, multi-scalar digital approach. The OpenAtlas system serves as both a semantic backend and a spatial database, enabling multi-domain data integration and networked querying. It was the underlying framework for the THANADOS project (Eichert, 2021; Eichert, 2024) and the data collection has been significantly extended for RELIC. Precisely in this stage of large-scale diachronic survey, differences in granularity, documentation regimes and category uses become most apparent, highlighting what often appears on the surface as gaps and anomalies in the data. This requires the implementation of a clear protocol for the harmonization and normalization of highly individualized records while also opening a possibility for their interpretative contextualization. In the case of “deviant burials,” the protocol is designed to keep descriptive observations (e.g., body position, treatment, placement) analytically separable from interpretive labels (e.g., “deviant”), so that contested categories can be compared across sites without detaching them from their evidential contexts. Spatial datasets in this workflow serve mainly as a common reference layer for contextual comparison. They support exploration of how burials labeled “deviant” occur within otherwise non-deviant cemetery environments, and how interpretation shifts when we compare them to local baselines and larger reference points.

Preliminary results and discussion

The existing dataset shows considerable regional variation in how Christianisation took shape materially (Vargha, Janovský, & Fajta, 2024). For instance, transitional cemeteries—sites where older burial customs merged with emerging Christian norms—often show spatial reconfigurations: changes in grave alignment, architectural interventions, or selective reuse (Parvanov, 2023; Čechura, 2024). These sites provide fertile ground for interpreting religious change as embedded in social practice and funerary negotiation (Vargha, 2022, 19-21; Štefan, 2024).

As the data collection for the south-eastern frontier zone of the Holy Roman Empire proceeds, the documentation landscape appears even more uneven. This phenomena is more than simply missing evidence, as it is the consequence of familiar foes like unsynchronized recording standards among different practitioners and national schools, limited research resources and differing levels of intensity in research, along with a disproportionate attention towards perceived power centres. A particularly sensitive part of the evidence concerns the earliest phases of site formation, where limited excavation and documentation, and selective publication practices often constrain what can be said out initial cemetery organisation and early church-cemetery relationships. These factors shape not only what is available, but also what is comparable: the same mortuary phenomenon may be described at different levels of precision, through different terminologies, and with different implicit interpretive assumptions.

This has direct implications for one of the paper’s central concerns: the interpretive category of “deviant burials.” Clear-cut thresholds in religious transformation are nearly impossible to detect, and, arguably, should not be expected. Instead, we observe multiple practices often framed as pre-Christian relics in popular behaviour, including church foundations within earlier cemeteries and burials reflecting different stages of religious transition.

Rather than treating “deviant” as a stable empirical class, preliminary harmonisation work shows that it frequently functions as an umbrella label in the mortuary record, covering a rich body of medieval evidence for glocal trends and contemporary scholarly biases (Milosavljević 2021; Parvanov 2024). How these temporally diverse factors play out in the formation of a narrative framed by the digital humanities will form one of the key discussions in our presentation. Initially a loan term from statistics, ‘deviant’ burial practices can misguidedly encompass all forms of bodily treatment. The juxtaposition of the strictly canonical expectations and the multitudes recorded in actual burial sites creates a disparity between grand historical narratives and recurrent archaeological variations.

In this sense, the preliminary results are less a claim about detecting clear-cut thresholds in religious transformation, and more a demonstration of how a DH infrastructure can support comparison without flattening controversy. By integrating the perceived gaps and anomalies in the data by recognizing and harmonizing differences in documentation regimes, descriptor vocabularies, and category usage into a common, inspectable framework, we can explore how interpretive labels (such as “deviant”) challenge broader concepts, such as the very idea of a linear Christianisation process (Thomas et al. 2017). The key discussion in our presentation will therefore focus on how digitisation and interface-driven comparison can bring together heterogeneous burial records to illuminate both glocal mortuary trends and the contemporary scholarly biases embedded in the categories we use to describe them.