Abstract

The logical system Hyperproof and the computer implementation of it--both created by Jon Barwise and John Etchemendy--present a radical new approach to modeling and teaching about reasoning. Hyperproof is a heterogeneous proof system that uses both sentences and diagrams as steps in proofs. This dissertation addresses important logical, philosophical, and pedagogical issues that Hyperproof raises. We formalize the syntax and semantics of Hyperproof, show that the major inference rules are valid, and give completeness results for four subsystems of Hyperproof. We address philosophical issues raised by the logical analysis, including the importance of the presence of homomorphism between diagrams and the worlds they represent. We also apply the constructivist educational philosophy to create a detailed explanation of how to teach analytic reasoning with Hyperproof, and we discuss how Hyperproof can be augmented to teach about reasoning with different types of domains