Résumés
Abstract
Discussions of dangerousness generally tend to be concentrated on the danger of violence in particular. The effective prediction of violent behaviour, if, indeed, it is the best guage of the objectivity of our knowledge of this phenomenon, is nonetheless an extremely difficult task. This article gives us an idea of some of the most current obstacles that confront the undertaking of prediction. Some additional thoughts follow, based on the author's earlier work on the subject.