Proof identity for mere mortals

Abstract

The proof identity problem asks: When are two proofs the same? The question naturally occurs when one reflects on mathematical practice. The problem understandably can be seen as a challenge for mathematical logic, and indeed various perspectives on the problem can be found in the proof theory literature. From the proof theory perspective, the challenge is met by laying down new calculi that eliminate ``bureaucracy''; techniques such as normalization and cut-elimination, as well as proof compression, are employed. In this note a new perspective on the proof identity problem is outlined. The new approach employs the concepts and tools of automated theorem proving and complements the rather more theoretical perspectives coming from pure proof theory. The practical approach is illustrated with experiments coming from the TPTP Problem Library.