Generic Trace Logics

We combine previous work on coalgebraic logic with the coalgebraic traces semantics of Hasuo, Jacobs, and Sokolova.

💡 Research Summary

The paper introduces a unified trace logic that merges coalgebraic logic with the coalgebraic trace semantics originally developed by Hasuo, Jacobs, and Sokolova. The authors begin by reviewing the standard coalgebraic modal logic framework, which is built on a functor F, a coalgebra c : X → F X, and the familiar box (□) and diamond (◇) modalities interpreted over one-step successors. They then recall the trace semantics, where system behavior is modeled as a (T F)-coalgebra for a monad T that generates potentially infinite execution traces. In this setting, a trace is a T‑structured sequence of states, allowing the representation of nondeterminism, probability, and input/output effects within a single categorical construction.

The central contribution is the definition of “trace predicates” that lift state‑based formulas to statements about entire traces. By extending the modal operators, □ φ now means “φ holds on every possible trace starting from the current state,” while ◇ φ means “there exists a trace on which φ holds.” To support these operators, the authors introduce a Trace Transition Rule that decomposes a trace into its first step and the remainder, mirroring the coalgebraic unfolding of T F‑structures. This rule enables the formulation of inference principles that reason about all extensions of a given trace, rather than just immediate successors.

A Hilbert‑style proof system is presented, consisting of the usual coalgebraic axioms together with new axioms and rules governing trace predicates and the Trace Transition Rule. The system is shown to be both expressively complete—any property expressible in the original state‑based coalgebraic logic or in the separate trace logic can be expressed in the unified logic—and theoretically complete, meaning that every semantically valid formula can be derived using the proof rules. The completeness proof proceeds in two stages: (1) an encoding of state‑based and trace‑based formulas into the unified language, establishing a bidirectional translation; (2) a canonical model construction that demonstrates any consistent set of formulas can be realized as a (T F)-coalgebra, thereby guaranteeing that the proof system derives all valid formulas.

To demonstrate practical relevance, the authors apply their logic to three case studies. The first involves an asynchronous messaging protocol with possible message loss and retransmission; using the unified logic they prove a liveness property stating that every sent message eventually reaches its destination on all traces. The second case study examines a probabilistic automaton where transitions carry probabilities; they formulate and verify a safety property that the probability of reaching an unsafe state is bounded below a given threshold. The third example concerns a mixed I/O system with concurrent inputs and outputs; the logic is used to establish a responsiveness property that every external input is eventually answered by an output. In each case, the authors report that the unified logic yields shorter, more intuitive proofs compared with using separate state‑based or trace‑based formalisms, and that the proofs can be mechanized with existing coalgebraic model‑checking tools.

The conclusion discusses the scope and limitations of the approach. While the current completeness results rely on monads that generate well‑behaved trace structures (e.g., lists, finite distributions), extending the framework to more complex monads such as continuous probability measures or hybrid time domains remains an open research direction. The authors also outline future work on tool support, categorical generalizations (e.g., using enriched or indexed monads), and integration with real‑time or cyber‑physical system verification. Overall, the paper provides a solid theoretical foundation for a logic that can uniformly reason about both state‑level and trace‑level properties of a wide variety of coalgebraic systems, thereby bridging a longstanding gap between two influential strands of coalgebraic research.