Transitive Closure and the Mechanization of Mathematics

We argue that the concept of transitive closure is the key for understanding finitary inductive definitions and reasoning, and we provide evidence for the thesis that logics which are based on it (in which induction is a logical rule) are the right logical framework for the formalization and mechanization of mathematics. We investigate the expressive power of languages with the most basic transitive closure operation TC. We show that with TC one can define all recursive predicates and functions from 0, the successor function and addition, yet with TC alone addition is not definable from 0 and the successor function. However, in the presence of a pairing function, TC does suffice for having all types of finitary inductive definitions of relations and functions. This result is used for presenting a simple version of Feferman’s framework FS 0, demonstrating that TC-logics provide in general an excellent framework for mechanizing formal systems. An interesting side effect of these results is a simple characterization of recursive enumerability and a new, concise version of the Church thesis. We end with a use of TC for a formalization of set theory which is based on purely syntactical considerations, and reflects real mathematical practice.

[1]  A. Avron Theorems on strong constructibility with a compass alone , 1987 .

[2]  S. Feferman Finitary inductively presented logics , 1994 .

[3]  Seán Matthews,et al.  Implementing FS0 in Isabelle: Adding Structure at the Metalevel , 1996, DISCO.

[4]  Yiannis N. Moschovakis,et al.  Abstract recursion as a foundation for the theory of algorithms , 1984 .

[5]  A. Avron On strict strong constructibility with a compass alone , 1990 .

[6]  Furio Honsell,et al.  A framework for defining logics , 1993, JACM.

[7]  Raymond M. Smullyan,et al.  Theory of Formal Systems. (AM-47) , 1961 .

[8]  Jörg Flum,et al.  Finite model theory , 1995, Perspectives in Mathematical Logic.

[9]  Seán Matthews A theory and its metatheory in FS 0 , 1994 .

[10]  Linda C. van der Gaag,et al.  THE LazyRMS: AVOIDING WORK IN THE ATMS , 1993, Comput. Intell..

[11]  Egon Börger,et al.  Trends in theoretical computer science , 1988 .

[12]  R. Smullyan Theory of formal systems , 1962 .

[13]  Hector J. Levesque,et al.  GOLOG: A Logic Programming Language for Dynamic Domains , 1997, J. Log. Program..

[14]  A. Dawar FINITE MODEL THEORY (Perspectives in Mathematical Logic) , 1997 .

[15]  Arnon Avron An exercise in an interactive geometrical research , 2005, Annals of Mathematics and Artificial Intelligence.

[16]  Alan Smaill,et al.  Experience with FS 10 0 as a framework theory , 1993 .

[17]  Emil L. Post Formal Reductions of the General Combinatorial Decision Problem , 1943 .

[18]  Alfred Tarski,et al.  Tarski's System of Geometry , 1999, Bulletin of Symbolic Logic.

[19]  Neil Immerman Languages which capture complexity classes , 1983, STOC '83.

[20]  D. Gabbay What is a logical system , 1994 .

[21]  J. R. Büchi On a Decision Method in Restricted Second Order Arithmetic , 1990 .

[22]  Ian A. Mason,et al.  Using typed lambda calculus to implement formal systems on a machine , 1992, Journal of Automated Reasoning.

[23]  Erich Grädel,et al.  On Transitive Closure Logic , 1991, CSL.

[24]  Frank Pfenning,et al.  The Practice of Logical Frameworks , 1996, CAAP.