Proof search has been used to specify a wide range of computation systems. In order to build a framework for reasoning about such specifications, we make use of a sequent calculus involving induction and co-induction. These proof principles are based on a proof theoretic notion of definition, following on work by Schroeder-Heister, Girard, and McDowell and Miller. Definitions are essentially stratified logic programs. The left and right rules for defined atoms treat the definitions as defining fixed points. The use of definitions also makes it possible to reason intensionally about syntax, in particular enforcing free equality via unification. The full system thus allows inductive and co-inductive proofs involving higher-order abstract syntax. We extend earlier work by allowing induction and co-induction on general definitions and show that cut-elimination holds for this extension. We present some examples involving lists and simulation in the lazy λ-calculus. Two prototype implementations are available: one via the Hybrid system implemented on top of Isabelle/HOL and the other in the BLinc system implemented on top of λProlog. © Springer-Verlag 2004.
CITATION STYLE
Momigliano, A., & Tiu, A. (2004). Induction and Co-induction in Sequent Calculus. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 3085, 293–308. https://doi.org/10.1007/978-3-540-24849-1_19
Mendeley helps you to discover research relevant for your work.