Department of Theoretical Philosophy
Predictivism in the Evaluation of Criminal Evidence
Some philosophers of science argue that a scientific theory which successfully predicts novel facts is ceteris paribus better supported than a theory which is built purposefully to accommodate those facts. In this thesis I examine the hypothesis that we should also be predictivists with regard to criminal evidence, i.e. all other things being equal, an explanation of the evidence in a criminal case is better supported if it successfully predicts certain facts than if it merely explains those same facts after they are known to the person constructing the explanation.
I argue that predictions can be evidence for two hypotheses: (i) The reliability of the reports a person makes, whether that person is an expert, witness or suspect, and (ii) the degree to which these explanations have been tested. I explore these two topics using Bayesian probability theory and reflect on the benefits of Bayesian modeling of the predictivist problem with regard to criminal evidence.
|Last modified:||26 October 2017 2.04 p.m.|