The installation La Loi normale des erreurs: les transactions Göring-Rochlitz constitutes the fifth development in a series initiated in 2014 concerning the issue of artworks looted from Jews during the Second World War.
Based on archival research, the piece visually translates the volume of trafficking that took place between March 1941 and November 1942 between the intermediaries of the dealer Gustav Rochlitz and Hermann Göring. On eighteen occasions, the two men "exchanged" paintings that had been seized from Jewish families with the goal of enriching, in one case, a personal collection, and in the other, a business stock.
The documentation arranged on the pedestal brings together the identification files for the paintings drawn up by the Einsatzstab Reichsleiter Rosenberg upon their arrival at the Jeu de Paume building in Paris, where the looted paintings were concentrated. These files, now available on online databases, have been completed and enriched with images and their current locations when the work has entered a public collection.
The installation at the Kunsthaus is the logical successor to the one presented at the Centre Pompidou, which focused on the holdings of the dealer Paul Rosenberg. Indeed, many works looted from the latter ended up serving as "bargaining chips" between Rochlitz and Göring. Having established this fact and seeing the significance of this historical aspect, I began searching for all kinds of documentation related to this dealer (lists, reports, interrogations, claim files from looted families, and iconographic sources) at the Archives Center for Diplomatic Affairs of the Ministry of Foreign Affairs.
In parallel with this "paper" archive research, I also rely on the primary database available on the internet (https://www.errproject.org/jeudepaume/card_advanced_search.php) to compile the complete listing of the exchanged works. This database, incredibly and unfortunately vast, remains at times incomplete and imprecise (paintings without images, inaccurate information, etc.). Regarding my interventions on this information, the idea is more to find the missing images than to correct errors in provenance.
Part of my research also took place, in an essentially intuitive manner, at the Library of the National Institute of Art History (INHA). It was, for example, in this institution where books are widely available in open access that I located the image of a Modigliani portrait missing from the database (https://www.errproject.org/jeudepaume/card_view.php?CardId=5979) by reading the latest work by Hans Christian Lohr; this type of discovery is added to the file.
The images that complete the files come from many different sources, sometimes from professional sites, sometimes from "amateur" sites fueled by volunteers or data scrapers. Catalogues, books, archival collections, and the internet (WikiArt, DHM, Lost Art, auction house websites, etc.).
Finally, the last intervention I permit myself is to indicate, when I find it, the current location of the paintings. This generally happens by combing through the online collections of museums; I also sometimes use image recognition software (this is how I located the Modigliani currently present—in a "not exactly legitimate" manner—at the Sprengel Museum in Hanover).
To summarize, the idea for each development of La Loi normale des erreurs, starting from a specific line of inquiry, is to elaborate a list of works and then put a "face" on each of them; it is then a matter of gathering them into a body of documentation and finally producing an installation that translates, piece by piece, the volume of the paintings in question.
Provenance errors—such as the one found for Renoir's Little Irene, to my knowledge the only one found among the paintings involved in the transactions—are left visible on the files even if I have identified them (https://www.errproject.org/jeudepaume/card_view.php?CardId=8996). Indeed, it is more important to me to put an image on a "registration number," for the sake of embodiment, than to correct a "detail": this task belongs, in my view, more to the historian or the database administrator than to myself.
