Computational and Experimental History of Science Workshop
Oxford, 2–3 June 2026 at Maison Française d'Oxford
Organisers: Jo Hedesan (Oxford) & Carmen Schmechel (KNAW)
Sponsors: Centre for Early Modern Studies (CEMS), Society for the History of Alchemy and Chemistry (SHAC), Oxford Centre for the History of Science, Medicine and Technology (HSMT)
On 2 and 3 June 2026, we convened a workshop at the Maison Française d'Oxford bringing together historians of science working with two rapidly developing methodological approaches: experimental history of science and computational history of science. Ten participants from several international institutions took part in person across the two days (afternoon of 2 June and morning of 3 June). The workshop was conceived as a dialogue between the computational and experimental approaches in view of a special issue of Notes and Records: The Royal Society Journal of the History of Science, which has been provisionally accepted.
The first day was focussed on experimental approaches. Peter Oakley (Royal College of Art) argued for reframing extended experimental reconstructions through design-development theory, using the ‘double diamond’ model of iterative and incremental development. By focussing on the topic of early modern furnaces, he argued that prototyping and iteration connect computational and experimental practice.
Thijs Hagendijk (Utrecht) presented on the pedagogical uses of historical experiment, arguing, with reference to hands-on reconstructions such as the ‘divine water’ recipe from the Leiden Papyrus, that direct laboratory experience gives students an understanding of historical alchemical processes that AI cannot substitute for. Questionnaire data collected from his students showed how their view of alchemy shifted once they had engaged with historical practices first-hand rather than only reading about them. However, the AI can be useful in furthering theoretical knowledge about the subject. Carmen Schmechel and Thijs Elfrink (KNAW) presented an initial computational analysis of the frequency and co-occurrence of fermentation-related terms across their assembled corpus of early modern Dutch and Latin sources, mapping how far concepts such as ‘ferment’, ‘yeast’ and ‘cheese’ cluster together across different text types.
The morning of 3 June was dedicated to computational methodology. Elisabeth Moreau (Palermo) described her approach of applying computational stylometry to the Paracelsian corpus. She highlighted the underlying commonality between scientific method and experimental approach that connects her work to the wider aims of the workshop.
Sarah Lang (Max Planck Berlin) reported on her ongoing work building a corpus of Processus Universalis recipes (17th century recipes connected to the Polish alchemist Michael Sendivogius). Her work has been focussed on tracing how the recipes changed and diverged over time. Guillermo Restrepo (Max Planck Leipzig) and Sarah Lang presented their approach to a combined pipeline of digital humanities extraction, quantum chemical modelling and wet-lab replication for reconstructing the pre-1800 ‘alchemical space’ using the EMLAP corpus and Ruland dictionary. Jo Hedesan, together with Vojtěch Kaše (West Bohemia) and Petr Pavlas (West Bohemia) presented ongoing work using the EMLAP corpus to disambiguate the multiple senses of the term ‘mercurius’ in early modern Latin alchemy.
The roundtable discussions centred on the ontology of computational versus embodied reproduction; the value of simple, shared prototypes as a common ground between computational and experimental work; the tension between embodied and textual knowledge, and the distinction between theory-driven and data-driven modelling. We also debated how much textual data is actually needed before a recipe can be reproduced computationally via quantum chemistry, how to disambiguate polysemous alchemical terms computationally, and the challenge of integrating the pre- and post-1800 chemical space.
Acknowledgements
We are grateful to CEMS, SHAC, and HSMT for their generous support, and to the Maison Française d'Oxford for hosting the event at no cost.
Report by the co-organiser, Jo Hedesan (Oxford)
