Search papers, labs, and topics across Lattice.
This paper addresses the failure of algebraic equality as a rewrite principle for measurement-bearing expressions due to the importance of observation reuse and definedness, particularly in the context of division. They introduce a unified semantics that tracks both provenance and definedness using token-sensitive enclosure semantics, leading to judgments for one-way rewriting and interchangeability. The work formalizes these concepts and theorems in Lean 4, demonstrating the limitations of equality in simplifying measurement-bearing expressions.
Algebraic equality breaks down as a rewrite principle for measurement-bearing expressions, revealing the subtle but critical role of observation provenance and definedness.
Ordinary algebraic equality is not a sound rewrite principle for measurement-bearing expressions. Reuse of the same observation matters, and division can make algebraically equal forms differ on where they are defined. We present a unified semantics that tracks both provenance and definedness. Token-sensitive enclosure semantics yields judgments for one-way rewriting and interchangeability. An admissible-domain refinement yields a domain-safe rewrite judgment, and support-relative variants connect local and global admissibility. Reduction theorems recover the enclosure-based theory on universally admissible supports. Recovery theorems internalize cancellation, background subtraction, and positive-interval self-division. Strictness theorems show that reachable singularities make simplification one-way and make common-domain equality too weak for licensed replacement. An insufficiency theorem shows that erasing token identity collapses distinctions that definedness alone cannot recover. All definitions and theorems are formalized in sorry-free Lean 4.