SCIENTIFIC REASONING IN ART

EVALUATING EVIDENCE IN PAINTINGS RESEARCH

To help clarify complex situations where the relative weight of evidence with respect to a given hypothesis needs to be established, it is possible to use Bayesian networks, probabilistic mathematical models. These networks can accommodate a large variation in data and can quantify the value or weight of each piece of evidence and how it influences a hypothesis in relation to other evidence. The networks are flexible and allow the incorporation of new evidence and thus quantify its influence on the overall chain of evidence. This project will apply Bayesian networks to questions regarding authentication of paintings, since these questions combine traditional humanities-based methods with scientific investigation using instrumental analytis. Keeping an overview of information delivered by different specialists and establishing its relative weight is a growing challenge. Bayesian frameworks will sharpen research methodologies by pointing to possible weak areas and will help researchers focus on relevant issues. Introducing Bayesian networks will show if the system can be of interest for a broader range of research questions and could ultimately lead to a fundamental change in research methodology.  

Van Gogh Museum director Axel Rüger and senior researcher Louis van Tilborgh reveal the newly discovered Van Gogh painting ‘Sunset at Montmajour’. All the technical and (art) historical evidence for attribution pointed in the same direction, but this is not always the case. Photo: AP/De Jongh, via www.theunderground.nl.

• • facebook
  • twitter
  • LinkedIn