Visualising argument structure

Argument mapping is the production of graphical maps of complex debates and ideas. The result - an argument map - is a chart presenting an overview of the reasoning. An argument map displays the evidential relationships among the propositions which make up some piece of reasoning or argumentation. Argument maps are useful in clarifying what reasoning actually is, communicating and evaluating complex reasoning. Argument mapping has been shown to help understanding and improve critical thinking. This thesis describes a new argument mapping system which supports multi-premise reasoning and true inference support. It utilizes a novel hi-tree based representation of the argument’s structure. An efficient algorithm for layout of hi-trees is given. Complexity results and proofs of correctness are provided. This thesis also describes a number of focus based interaction techniques for visualisation of argument maps, and a number of different layout styles which provide more compact layouts of larger argument maps. Fast algorithms for layout of hi-trees and layered trees based on these styles, which now form the basis of a commercial argument mapping application, are given. Lastly the use of Quadratic Programming (QP) as an alternative approach for horizontal positioning of hi-trees is explored. Layouts produced with QP are compared and evaluated against the layouts produced by specialized algorithms.