Erased from the playbill: how AI is helping to map Nazi censorship of German music theatre

A research project led by OCAD U Associate Professor Dr. Helmut Reichenbächer is examining how Nazi ideology reshaped what German stages performed, using artificial intelligence (AI) to transform some 6,000 pages of historical theatre listings into open data.

In a little more than a decade, the German stage lost a substantial part of its music-theatre repertoire, along with many of the people who had made it.

“More than half the operetta performances on German stages in the final Weimar years featured works by composers whom the Nazi regime would soon defame and drive from the stage. Behind every title that vanished from the playbills is a career destroyed or a life in danger. Our data makes that loss visible at scale and our archival research will recover the personal stories behind the numbers," says Dr. Reichenbächer.

The research project team will analyze over 500,000 performance records from more than 900 German-language stages, drawn from printed sources such as the Deutscher Bühnenspielplan and the Deutsches Bühnenjahrbuch.

He developed an interest in opera and other forms of music theatre growing up in Frankfurt am Main, Germany. It stayed with him through his studies at the Hochschule für Musik Köln and the University of Cologne, and when he obtained a PhD in English at the University of Toronto. Before joining OCAD University, he was editor of publications for the Canadian Opera Company and associate producer in Radio Music at the Canadian Broadcasting Corporation (CBC).

Join Dr. Reichenbächer for the sabbatical presentation, “Performing Under Pressure: Music Theatre and the Effects of Nazi Censorship” on Tuesday, October 28, from 3 to 3:45 p.m. at 230 Richmond St. W., room 322. 

OCAD University caught up with Dr. Reichenbächer about his current research project, “From Weimar to Nazi Germany: A Digital Humanities Investigation of Music-Theatre Censorship,” funded through an Insight Grant from the Social Sciences and Humanities Research Council of Canada (SSHRC).

What is the focus of this research project, and what inspired you to explore this topic?
I am studying how cultural censorship reshaped the programming of music theatre: opera, operetta, and incidental music. The project traces the shift from the diverse and experimental programming of the Weimar Republic (1918–1933) to the ideologically restricted repertoire of the Nazi era (1933–1944), which ended with the closure of the theatres.

At the centre of the work is a digital pipeline that we are building to draw material out of historical print sources and convert it into a machine-readable dataset. Optical character recognition and AI do the extraction, and every stage is checked by hand against the printed originals: a dataset of this kind is only as good as its error rate.

This makes two kinds of analysis possible at once: the wide view of how repertoire shifted across some 900 German-speaking theatres inside and beyond the Reich, and the close view of what was staged in a single house in a single season.

The approach has already been tested. Under an earlier SSHRC Insight Development Grant, I aggregated performance data by hand for a single targeted work, Mendelssohn’s incidental music to A Midsummer Night’s Dream; the resulting study appeared in the Journal of the American Musicological Society in 2025 and led me directly into archival research. This project automates that aggregation and scales it up.

What drew you into this field of study in the first place?
Two developments brought the subject into focus. The first is the wave of political polarization now visible in many parts of the world, which has made questions about who controls cultural programming feel considerably less like history. The second is the pace of change in digital technology. Print sources that would once have taken most of a career to compile by hand can now be assembled into a structured dataset within a few years.

What question is your research trying to answer, and why is this important?
The dataset we are building will support analysis across a wide range of parameters: which composers thrived and which were suppressed, which theatres kept defamed repertoire on the bill for a while, and which of them held out longest.

Alongside that broad picture, I am studying the individuals who made the programming decisions and the professional careers they built or lost in the process. The dataset tells me where to look: which case studies merit closer attention, and how particular people operated within an increasingly pressurised political environment. That is the entry point for my searches in German archives, including the National Socialist bureaucratic records held at the Bundesarchiv in Berlin.

How do you see the research findings contributing to affecting people’s lives in the real world?
Cultural censorship continues to constrain creativity, artists’ livelihoods, and in some places artists’ lives. That makes it all the more important to examine how censorship actually operated in the past, in detail rather than in outline.

The historical record documents conformity and resistance alike, along with the costs of each, and it gives us a sharper instrument for recognising the same mechanisms at work in the present.

Everything we produce will be openly available, published as linked open data through LINCS at the University of Guelph and mirrored at the Saxon State and University Library in Dresden, so that others can put questions to the material that have not occurred to me.

Are you working with any collaborators, institutions, or funding bodies on this project?
Two co-applicants anchor the Canadian side of the work. Professor Isabel Meirelles, an award-winning information designer in the Faculty of Design at OCAD University, turns patterns that exist in the data into forms a reader can see and interrogate. Professor Dr. Susan Brown, Canada Research Chair in Collaborative Digital Scholarship, Tier I, at the University of Guelph, leads LINCS, a national platform that publishes cultural datasets as linked open data. She and her team advise on how our material is modelled, using the international standard for cultural heritage data (CIDOC CRM), and LINCS will publish and maintain the dataset.

In Germany, Dr. Christian Kämpf coordinates the Specialised Information Service for Musicology at the Saxon State and University Library in Dresden, where he is responsible for musiconn.performance, an infrastructure service for cataloguing and publishing musical events for research.

Since 2025 we have worked on encoding historical performance records so they can be linked across institutions rather than isolated in individual projects. Kämpf and his team, members of the NFDI4Culture consortium for cultural heritage data, will align our records with the Gemeinsame Normdatei (GND), the authority file that German libraries and archives depend on, and Dresden will host a permanent mirror of the published dataset.

Dr. Ulrich Menzefricke, Professor Emeritus of Operations Management and Statistics at the Rotman School of Management, University of Toronto, advises on the statistical analysis of the corpus. Pouyan Shahidi, a digital humanities method specialist and doctoral candidate at Indiana University Bloomington, maps performances across a theatre landscape whose borders and administrative boundaries shifted repeatedly across the period, which shapes almost every regional comparison we can make.

My research assistants Beatrice Hunter, a graduate of the University of Toronto’s Faculty of Information, brings rigorous attention to data management, and Zhiqi Gao, a graduate of York University’s Master of Artificial Intelligence Management, is central to developing the automated aggregation pipeline that extracts structured performance data from print sources such as the Deutscher Bühnenspielplan. Students working alongside them, will acquire skills in AI-assisted data extraction, semantic modelling, and archival research that transfer well beyond this project.

Last but not least, Johannes Debus, music director of the Canadian Opera Company, serves as artistic advisor. He brings a practitioner’s reading of the repertoire: what these works demand of a company, and what a house signals when it programs such a work or quietly takes it off.

What do you hope people will take away from your research?
I hope people will come away thinking about individuals and the range of their responses to political and cultural pressure, from open opposition through quiet manoeuvre to full conformity. Institutions do not program a season; people do, and they do so under specific political conditions. That was true of German theatres in the 1930s and 1940s, and it remains true of cultural life today. That is why the project will end in more than a dataset: a public web exhibition, “Narrativizing Data,” will set those individual stories back alongside the numbers.

This research project is supported by the Social Sciences and Humanities Research Council of Canada, Insight Grant.