Kripke-Platek set theory (KP) is a foundational system in mathematical logic that describes a restricted universe of sets. Proof theorists want to understand exactly how strong this system is, meaning which mathematical statements it can prove and how far its reasoning reaches. Over the decades, two separate technical frameworks were developed to do this kind of analysis: one called operator-controlled derivations, which tracks how proofs are built up step by step using labeled operations, and another due to the logician Jean-Yves Girard, which uses structures called dilators to measure the size and complexity of proofs in a more abstract, categorical way. These two approaches were powerful but looked quite different on the surface.
The paper bridges these two frameworks by recasting the operator-controlled analysis of KP in the language of category theory, specifically using the concept of functors, which are structure-preserving maps between mathematical categories. By doing this, the authors show that the two approaches are not really competing but are instead two views of the same underlying phenomenon. The translation makes the operator-controlled method more conceptually transparent and connects it directly to Girard's dilator-based setting.
As a bonus application, the authors give a new proof of a result known as Girard's boundedness theorem. This theorem concerns functions defined over a particular level of the constructible universe of sets, specifically the level reached after a certain number of steps indexed by the Church-Kleene ordinal. The theorem says that any such definable function from that ordinal to itself cannot grow faster than some function describable by a recursive dilator. The new proof, made possible by the unified framework, is cleaner and more illuminating than previous approaches.