Software Engineering Institute

Assurance cases provide a structured method of explaining why a system has some desired property, for example, that the system is safe. But there is no agreed approach for explaining what degree of confidence one should have in the conclusions of such a case. This report defines a new concept, eliminative argumentation, that provides a philosophically grounded basis for assessing how much confidence one should have in an assurance case argument. This report will be of interest mainly to those familiar with assurance case concepts and who want to know why one argument rather than another provides more confidence in a claim. The report is also potentially of value to those interested more generally in argumentation theory.