Statistics
L’épuration professionnelle du monde du spectacle à la Libération. Histoire et sciences mathématiques
Publié le - Les Annales. Histoire, sciences sociales
Abstract Written by a historian and a mathematical scientist, this article explores the epistemological stakes of interdisciplinary collaboration by focusing on a specific case study: the purge of performing artists after the liberation of France. While the artists most compromised during the occupation were brought before purge tribunals, less serious cases were referred to specialized commissions comprised of their peers. In any legal proceeding, it can be hard to reach a verdict when it comes to questions of fairness or potential discrimination, and this was especially true for these purge commissions. The authors show how mathematical formalism, while obviously not replacing historical inquiry, can extend its reach, offering multiple ways to apprehend an elusive reality thanks to the versatility of an abstract language. Progressing step by step through the modelling of the question and the analysis of the data, they explain the increasingly complex statistical and mathematical approaches mobilized to observe forms of jurisprudence that escape more traditional analysis—arriving at the innovative proposal to treat a trial involving human decisions as a complex algorithmic process. Adapting a causal inference approach designed to evaluate the fairness of “black box” type algorithmic processes brings to light unprecedented results, hitherto hidden in the data. These findings, in turn, lead to new insights for the historian.