Dialectic Evidence Assembly for Discovery

We propose and demonstrate a dialectical framework for the assessment and assembly of evidence in discovery processes. This paper addresses the stimulation and capture of dialectical argumentation in the context of complex (wicked) problems. Holonic representation is developed to structure processes hierarchically for the modeling of complex problems. Interval Probability Theory (IPT) is modified to produce an evidential reasoning calculus to represent dialectical argument through the separation of evidence for and evidence against a hypothesis. Support for and against any hypothesis can then be assembled through a weighting of the relevance, importance and degree of evidence. Uncertainty surrounding any hypothesis can be ecomposed into randomness, vagueness, conflict, incompleteness and relevance and can be managed within the framework. The framework is llustrated with real examples of discovery from two energy related complex (wicked) problems.