Cultural heritage institutions for performing arts face the problem that their subject is intangible and that any object that can be collected, documented, and preserved is of second order. A drawing might sketch a stage design, a photo or video might document a specific performance, but the actual work exists only in the moment of its enactment. Thus, there is no unequivocal method of how to preserve performing arts.
When it comes to developing an order for such objects, the situation is different for museums and archives. In museums, such objects are upgraded by placing them in specific narratives that also provide them with a higher degree of self-sufficiency. In archives, the context is primarily one of provenance – objects or rather records derive from a specific person or company, and it is through these actors that they are also connected to the works of performing arts. It is a central aim of the project presented here to provide additional perspectives on these archival holdings and to better understand works and their contexts.
Context
The three original archives that since 2017 have merged into the Swiss Archive of the Performing Arts (SAPA) had all developed their own methods to document performing arts. The former Theatre Collection in Bern had been closely aligned with the University of Bern’s Institute for Theater Studies and was distinguished by its efforts to collect information in the form of theater programs and newspaper clippings and thereby documented a canon of major Swiss theaters and their productions.
The Swiss Dance Archives in Lausanne had a more community-orientated approach, which focused on individual choreographers and dancers and also collected objects like costumes that were part of their estates. The Dance Media Library in Zurich took its starting point from a specific media practice, i.e., the fact that more and more dancers and choreographers documented their productions and rehearsals on video. As video is a fragile form of documentation, it quickly became apparent that an institution was needed to preserve the recordings and make them available.
The three different approaches have led to distinct archival practices that included specific databases and came with inherent data models. One of the main tasks since the merger four years ago has been to create one unified database containing information about the holdings at all three locations and to connect them in a meaningful way.
Due to the earlier merger of the archives in Lausanne and Zurich, they already had a shared database, which was structured according to traditional archival requirements, i.e., it primarily documents the provenance of records in hierarchical order. Such a solution falls short when it comes to domain-specific information. In our case, this means that the records such as films, videos, and photographs document performances, and information about the documented works including contributors, venues, and original works was stored in additional free text fields. This, in itself, made it difficult to unambiguously identify these entities. And the fact that the spelling of foreign entities can differ in French and German did not make the situation easier.
The approach in Bern differed as there was not a single archival database but many different ones, from elaborate spreadsheets to many FileMaker Pro databases. All these databases followed one specific purpose and contained primarily domain-specific fields. There are individual databases for stage designs, masks, theater producers, and theatrical productions. The latter contains detailed information about more than 60,000 productions shown in Switzerland since the late 19th century.
Hybrid modeling
To be able to store all this information in one common graph database, we worked with the Bern University of Applied Sciences to outline a data model based on CIDOC CRM, its extension FRBRoo, and the not yet finished archival standard Records in Context (Estermann, Schneeberger 2017).
Here we could also build on what others have done over the last two decades. One of the earliest successfully implemented models is found in AusStage, a database of theatrical productions in Australia, which later was further developed with IbsenStage in Norway. Both projects connect evident classes such as works, events, contributors, venues, etc. in a traditional relational database approach. They are neither connected to archival holdings nor other databases or authority files, but they serve and fulfill one specific purpose (Bollen 2016).
Since then, there have been several approaches based on semantic modeling. Between 2010 and 2013, the European Commission funded the European Collected Library of Artistic Performance project (ECLAP), which aimed at aggregating data from various European sources and developed a model aligned to the Europeana Data Model (EDM) (Bellini, Nesi 2013). The Carnegie Hall in New York provides its data reusing well-known ontologies like FOAF or Schema.org. The German Centre of the International Theatre Institute (ITI) in Berlin has taken the opposite approach and developed its own domain-specific ontology, the Performing Arts Archive Model (PAAM).
At the same time, there has been a discussion of the usage of first FRBR and then FBBRoo, the extension of CIDOC CRM that is primarily intended for bibliographic information but also provides classes and properties for performing arts. While this discussion can be traced back to a suggestion by Patrick Le Boeuf at the SIBMAS congress twenty years ago, I am not aware of any implementations with actual data.
In 2008, Martin Doerr, Patrick Le Boeuf, and Chryssoula Bekiari (2008) demonstrated how the well-known FRBR axis with work, expression, manifestation, and item looks like for performing arts and how CIDOC CRM provides the chance to relate the different stages of concretization to the artifacts in museum collections and archives. One might argue that the perspective presented here actually does not connect the works of the performing arts with related artifacts but the other way around. It extends the annotation of the objects and contextualizes them but does not reverse the hierarchy between material and immaterial entities because the specific activities that characterize the performing arts are missing. When Le Boeuf (2012) presented advanced modeling examples with activities a few years later, they were also strictly related to the artifacts’ production and not the performances themselves.
Implementation
For our data model, a more detailed rendering of the productions was necessary, and activities were represented as properties adopted from the RDA Registry (e.g., “has choreographer”, rdau:P60433). With the ongoing implementation of the model there have been two significant changes: to allow for a higher degree of compatibility, instead of own classes, properties, and attributes, we are now using the original ontologies. We combine these with our own controlled vocabularies and document the implementation in an evolving profile as the migration of the different legacy databases progresses.[1]
The second change is that we reified the activities so that we can represent them more accurately. So, instead of saying that a production has a specific stage designer, by using CIDOC’s Activity class, we can classify the activity according to our controlled vocabulary and document that it was designated differently in the program, e.g., as an “installation.” (Weiberg 2020).
Outlook and conclusion
For a future digital ingest process, we plan to extend the profile with PREMIS, the Preservation Metadata Implementation Strategies. Therefore, our data can be divided into three spheres or perspectives: the core archival information will be described with RiC, the technical metadata for the preservation of media items will be documented with PREMIS, and the domain-specific information regarding the absent performing arts works with CIDOC CRM plus FRBRoo. Even if we do not use the latter two ontologies for all our data, they play a central role in capturing our holdings and allowing to take multiple perspectives on the documents and objects.
A first public version of our database performing-arts.ch is available with some but by far not all of our data. More data and updates to the interface will be added over time.
[1]http://sapa.github.io/spa-specifications
References
Bellini, P., P. Nesi (2013) A linked open data service for performing arts. In P. Nesi, R. Santucci (eds.), Information technologies for performing arts, media access, and entertainment: Second international conference, ECLAP 2013, pp. 13–25. Berlin, Heidelberg: Springer.
Bollen, J. (2016) Data models for theatre research: People, places, and performance. Theatre Journal 68/4), pp. 615-632.
Doerr, M. et al. (2008) FRBRoo, a conceptual model for performing arts. In 2008 Annual Conference of CIDOC, Athens, pp. 15-18.
Estermann, B., C. Schneeberger (2017) Data model for the Swiss performing arts platform. [Available at: https://datahub.io/dataset/spa-data (Accessed: 1 April 2021)].
Le Boeuf, P. (2012) Towards performing arts information as linked data?. In SIBMAS 2012 Conference. [Available at: https://hal-bnf.archives-ouvertes.fr/hal-00807942. (Accessed: 1 April 2021)].
Weiberg, B. (2020) Modeling performing arts: On the representations of agency. Arti dello Spettacolo/Performing Arts 6/6, pp. 50–56. [doi: 10.17613/6tsm-4787].
