Lukas Sebold, Leonie Aschmann: Report on the conference – “Data for History 2021” – Day 3

Session 6 – Uncertain Time and Space

In the first presentation of the day, ”Chronology Statements for nodegoat: a temporal topology to interface with vague, relational, and actionable dates”, Pim Van Bree and Geert Kessels (LAB 1100, The Hague) introduced the platform Nodegoat. Nodegoat is a web-based research environment for the humanities, providing services to create e.g. network, linked and spatial data. However, there is also a need for complex statements to express temporal uncertainty.
The speakers pointed out that historians in general lack vocabularies to represent temporal uncertainty. This is reflected in common implementations in DH that miss flexibility and a standard for relational temporal statements. Hence, it is difficult to model some event happening before another without referring to a specific date. To implement a humanly and computationally interpretable temporal topology, Nodegoat relies on ChronoJSON. ChronoJSON ensures computational efficiency and fits into the dynamic interface of the platform.
Temporal statements on events in ChronoJSON are split into beginning and end, which in turn are split again. This allows to model the start and end of events as a range. The ranges themselves can be described in relational terms and each statement in them is referenceable. ChronoJSON also allows for specifying custom cycles and periods, like university semesters. This implementation also enables researchers to filter data.

In the ensuing discussion two questions came up. The first expressed concern regarding the implementation’s complexity and possible hesitancy among researchers. The creators of the platform admitted a steep learning curve for researchers but pointed out that, in their experience, the benefits ultimately win over many researchers. Additionally, the creators offer to work with interested scholars and help them to navigate the platform.
The second point raised referred to the possibility of contradictory temporal statements when it comes to relational data and the possibility to automatically detect them, for example through logical reasoning. According to the speakers, it is at least possible to define illogical statements on the platform, so that the researcher can be notified if a statement is recorded that meets these criteria. However, there is no standard on the platform for what is considered illogical. This can only be determined by the individual researcher.

In the second part of the session, Grzegorz Myrda and Tomasz Panecki (Polish Academy of Sciences, Warsaw), gave a talk on “The problem of using persistent identifiers for historical geographical objects”. The researchers showed that there are four constitutive properties of geographical objects which allow identification over time: name, location/geometry, type of object, and mereological relations. In a historical context, all four of these may be subject to change, making it difficult to model them in a computational setting.
One challenge here is the tension between the practical question of “what works” and the philosophical question of the identity of objects over time. Harmonization can be accomplished by selecting a reference data set that uses permanent identifiers (PID). In a historiographical context, this means to choose the right identifiers for the right kind of object, which has to be set apart from its actual real world features at any point in time.
Another problem with common standards is the existence of a registration authority which may or may not be accessible and flexible enough for researchers’ purposes (for example: URI-URN). One further challenge arises from working with governmental data, which is confined to present-day borders often differing from their historical counterparts. The researchers proposed WikiData as their preferred imperfect solution.

Following the presentation, Grzegorz Myrda and Tomasz Panecki clarified the distinction between geographical entity and place: the latter is only defined by coordinates, while the former describes a human conceptualization which can have a fuzzy location or even change it over time.

Session 7 – Space and Agents

Philipp Schneider’s (Humboldt-Universität zu Berlin) talk about “Heraldry as a Historical Source to Conceptualize Medieval Spaces and Agents” was concerned with modelling the relations between agents, places, and events and the importance of context for their interpretation.
The main difficulty in his case is to avoid the ambiguity that arises from the fact that the same person may be associated with different places in different capacities, so that no unique place can be assigned to him or her but rather depends on the context of the mention in the source. For example, the coat of arms of the same person may be mentioned several times in the same collection of coats of arms, each time in reference to a different regional overlord or to different geographically localised events such as tournaments.
Schneider’s solution to this ambiguity is to introduce a new class, “role”, which specifies the relationship between agents and places, depending on the respective setting. This was achieved by integrating the class “Role” into the data model, reusing classes from CIDOC CRM (mainly “E17 Type Assignment”). In this way, the interpretation of the role that a person has filled in a particular context is still a matter of interpretation, but part of the model itself. In CIDOC-CRM, the role is given via the type assignment and allows for connection to any sources on which it is based.

The subsequent discussion focused on both the technical solution as well as the historical meaning of roles. Regarding the technical solution, Francesco Beretta asked whether an alternative ontology like SCHEMA might be preferable to CIDOC-CRM, in this context. An alternative could be to treat roles as temporal entities. When answering the question about the technical implementation, Schneider pointed out that he separates the role itself and the assignment of said role, which gives the model more flexibility.  
Furthermore, the status of what the “roles” actually are was of interest. Here Schneider considers them only as research concepts, not as historical objects themselves.

In the last presentation of the day, “Space, time and period – the Neoclassica approach”, Simon Donig (Universität Passau) introduced the project’s aim to develop appropriate instruments for research in the area of the history of aesthetic culture.
To enable an analysis of material culture of interior spaces and to model objects and their environment, the Neoclassica project implemented a research oriented ontology. This ontology, integrated with CIDOC CRM, accounts for complex representation of actors, space and time. Thereby, the Neoclassica ontology allows for modelling and formalizing concepts like periods in art history in relation to time and space. It also enables the mapping of such complex relations as the interplay of various agents and workshops, involved in the creation of artefacts, represented in the ontology. This is made possible by the introduction of a new class “workshop”, that connects artists and artisans to artefacts as well as places of origin.

In the ensuing discussion Simon Donig stressed the importance of transparency when publishing models for representing temporal and spatial concepts. Documentation should not just show the results of research and implementation, but disclose the underlying processes. The researcher stated the advantages of CIDOC CRM to model temporal uncertainty, by being able to represent vagueness and different periodization and express overlapping periods.

 

---

About the authors:

Leonie Aschmann is currently studying for a master’s degree in history with a focus on modern history at the Humboldt-Universität zu Berlin. Previously, she completed a bachelor’s degree in media and communication studies and political science.

Lukas Sebold has a master’s degree in philosophy with a focus on political philosophy and is now studying for a master’s degree in history with a focus on Digital History at the Humboldt-Universität zu Berlin.

---

---