Making the Bayesian approach to criminal evidence work: from problems to practice.

In criminal trials mistakes are sometimes made when reasoning about the facts. Bayesian approaches provide a promising mathematical framework which can support evidential reasoning and prevent errors. However, current forensic applications run the risk of ignoring that legal proof is not just mathematics but also applied philosophy and it is bound by legal-procedural constraints. This research aims to develop philosophically grounded Bayesian methods for legal evidential reasoning using ideas from the philosophy of science. It will investigate how these methods can conform to (Dutch) legal proof standards, in particular the presumption of innocence and the right to remain silent.