In reductive proof search, proofs are naturally generalized by solutions, comprising all possibly infinite structures generated by locally correct, bottom-up application of inference rules. We propose an extension of the Curry-Howard paradigm of representation, from proofs to solutions: to represent solutions by possibly infinite terms of a dedicated lambda-calculus. This new, comprehensive approach to proof search is exemplified with the sequent calculus LJT for intuitionistic implication logic. A finitary representation is proposed, comprising lambda-terms extended with a formal greatest fixed point, and a type system that can be seen as a logic of coinductive proofs. In the finitary system, fixed-point variables enjoy a relaxed form of binding that allows the detection of cycles through the type system. Formal sums are used to express alternatives in the search process. Moreover, formal sums are used in the coinductive syntax to define "decontraction" (contraction bottom-up) - an operation whose theory we initiate in this paper. As a semantics, we assign a coinductive lambda-term to each finitary term. The main result is the existence of an equivalent finitary representation for any full solution space expressed coinductively. This result is the main ingredient in the proof that our logic of coinductive proofs is sound and complete with respect to the coinductive semantics. These results are the foundation for an original approach to proof search, where the search builds the finitary representation of the full solution space, and the a posteriori analysis typically consisting in applying a syntax-directed procedure or function. The paper illustrates this for proof search and inhabitation problems in the simply-typed lambda-calculus, reviewing results detailed elsewhere, and including new results that obtain extensive generalizations of the so-called monatomic theorem.
翻译:在反感应证据搜索中,证据自然被解决方案普遍化,包括由本地正确、自下而上地应用推断规则产生的所有可能无限的结构。我们提议扩展Curry-Howard代表模式,从证据到解决方案:代表解决方案,可能以专门的羊羔计算仪的无限条件。这种新的、全面的检验搜索方法以直观含义逻辑的顺序缩略图LJT为示例。提出了一种有根有根有根的表述,包括由正式的固定点延长的羊腿术语,以及可以被视为硬币证据逻辑的分类系统。在固定系统中,固定点变量享有宽松的束缚形式,允许通过类型系统探测周期。在搜索过程中,正式的金额被用来表达替代物。此外,在硬币合成合成词中,正式的金额用于定义“违约” (合同完成的底部),这是我们在本文中启动的一种理论, 作为一种语义,我们给每个固定证据的逻辑分析结果加上了一种直截面的直截面、直截面的直截面分析结果。