A new digital platform, The Minutes of Multilateralism (MoM), was recently launched as a pilot project by the CDHM. Where did the idea for the platform originate?

Grégoire Mallard (GM): In late 2024, initial discussions took place between archivists and researchers at the Centre on how the CDHM could facilitate a better integration of various International Organisations’ (IOs) digitised archives that had begun to be published online. The IOs fortunate enough to benefit from budgets to digitise their archives had placed PDFs on their own websites, but this siloed approach to the memory of multilateralism didn’t allow researchers and the general public to look at the complex history of multilateralism at an aggregate level. The goal then was to carefully select different archives of multilateralism and make them searchable in relation to one another with the help of visualization tools. Instrumental to this plan were associate professor of computational social science at the Engineering School in Computer Intelligence (EPITA) and CDHM advisor, Florian Cafiero, and professor emeritus and Director of the University of Helsinki Centre for Digital Humanities (HELDIG), Eero Hyvönen, who helped the CDHM develop an idea to select only those documents that looked like parliamentary debates: minutes of board meetings or general assembly meetings of IOs. This premise emerged from a pioneering semantic web project developed by Eero and his team, which made the corpus of Finnish parliamentary debates accessible to the public and had subsequently gained visibility by raising accountability in Finnish politics. Emulating this approach, the prospective Minutes of Multilateralism platform would give the global public a view on the ‘parliamentary’ discourses of world diplomacy in the form of general assemblies and the executive bodies of IOs. Our existing partnership with the UN Library & Archives in Geneva provided a natural starting point: having already digitised all the debates of the League of Nations General Assembly, they authorized the MoM platform to place the corpus in a common pool of archives. This process meant abandoning the image format of PDFs and instead adopting a textual format to search all words while filtering text according to desired speeches, topics, and tags for individual speakers – including the biographical data of historical figures. With an assembled team of talented historians, semantic web researchers, computational sociologists and archivists, the current prototype took shape using tailored tools, such as extractable OCR text, named entity recognition, and knowledge graphs. 

Given the volume of minutes, differing formats across organizations, and the breadth of historical figures included - this is an ambitious project. How do the CDHM and its partners navigate the project scope? 

Davide Rodogno (DR): The project is ambitious and in fieri. Its nature ought to be kept in mind by all users, first and foremost by ourselves. We, researchers and scholars in digital humanities, had to design an open research tool from the outset, one that is constantly revised, with mistakes – and there are many – continuously corrected. Our goal was to develop a pilot platform as a basis for exploring the larger CDHM vision of a repository linking archival collections across IOs to more easily find and follow connections among individuals and topics forming the breadth of historical material on the birth of modern multilateralism. We believed we had to publicly acknowledge the advantages and risks of compiling such an unprecedented set of minutes from different organisations and an equally unprecedented breadth of historical figures. For instance, in offering an accessible user interface to the public in the service of bolstering engagement with the history of multilateralism, we must also be open to the possibility that some biographies are inexact from the onset. As such, the MoM is not a finished product, but rather a jumping-off point for the continued realisation of shared goals, in scope and spirit, to combine the algorithmic with the heuristic. It is our responsibility to communicate the risk of perpetuating historical mistakes in our sources, including their blind spots and biases, while also constantly striving to correct them. Sharing the limits of the project and showing its innovative nature and its possibilities is the only way I know to navigate the MoM project. Had we set a different criterion, like technical and historical perfection, we would have had to wait indefinitely before publishing the prototype and presenting it to the public. We preferred to propose a new tool quickly, and let our users help us to correct its mistakes.

What might a user expect from interacting with the MoM tool? 

GM: Whereas before, a researcher (myself included) would need to download potentially hundreds of individual PDFs across different organisations and collections to generate and investigate research questions, the MoM is a proof of concept, starting with the UN Library & Archives’ digitised minutes of the League of Nations and those of the General Assembly of the United Nations, that uses an OCR-text-based linked open data (LOD) service and semantic portal to support dynamic, interdisciplinary digital humanities research across multiple texts, fonds, historical figures, places, and dates simultaneously. It acts as a thematic entry point for any user – whether they’re interested in a particular historical figure, in speeches on specific topics, in understanding the development of concepts in multilateralism across time and space – to navigate and engage with corpora of minutes in innovative ways. Crucially, the user is then able to discover, access and return to the original PDF documents directly from our IO partners’ archival collections – which provide a formal context and environs for the documents. So, while the minutes may not all be newly accessible to the public, the filters and lenses allowing a user to locate interconnected and previously unrelated connections as an instrument of research are – thanks to cutting-edge tools that helped create the platform. And as time passes, the user should expect the collections to be enriched with new material from other organizations with which we are now working to enrich the portal, like the International Parliamentary Union (IPU) and many others, in International Geneva and beyond.

The MoM facilitates transversal research methods that are increasingly important across Humanities disciplines; how can tools like this support evolving research priorities?  

DR: The MOM is a tangible example of how research evolves. The project’s nature invites researchers to explore other ways of collecting data and interpreting sources, to connect them differently. For example, for historians like me, used to visiting archives and working on a box, a folder, a dossier after another, the digital visualisation of hundreds or thousands of documents on a single individual, entity, place or time invites us to think differently. The combinations offered by the MOM allow scholars to try, at no cost to them, and fail, to rethink and visualise sources assembled for the first time, differently. That’s the magic of digital humanities. It is not a panacea, and we are certainly not preaching for scholars to give up their disciplinary identities. The MOM is a tool that shows, in a tangible way, that a transversal research method can add a different, new perspective; in the end, it enriches how we research and the research itself.   

What can we look forward to in the coming years regarding the MoM project? 

GM: There are several ways the MOM project is already contributing to collaborative offshoots with other CDHM workstreams. Two of our Visiting Fellows, Florian Cafiero, who was at the origins of the project, and Jean-Philippe Cointet from the médialab at Sciences-Po Paris have incorporated it into their coursework for Geneva Graduate Institute students, for example. We have plans in the future to develop trainings with the tool in mind that can create uptake on digital humanities and computational methods in research. We also have events and presentations involving the tool in the pipeline for the coming year. Other exciting projects related to the MoM include the CDHM’s work on a DiploBERT and an agentic Retrieval Augmented Generation (RAG) system, which will aid in the CDHM’s developing efforts to understand how AI might be used to analyze archives, improve named entity recognition, create metadata, and more.