Kripke Models for Classical Logic

We introduce a notion of Kripke model for classical logic for which we constructively prove soundness and cut-free completeness. We discuss the novelty of the notion and its potential applications.

💡 Research Summary

The paper introduces a novel Kripke‑style semantics tailored for classical logic, addressing the long‑standing mismatch between traditional Kripke models—originally designed for intuitionistic reasoning—and the law‑of‑excluded‑middle and double‑negation elimination that characterize classical systems. The authors begin by reviewing the standard intuitionistic Kripke framework, emphasizing its reliance on monotonic truth propagation along an accessibility relation and its inability to accommodate classical tautologies without collapsing into triviality. To overcome this, they propose a two‑layered world structure: regular worlds that behave like conventional Kripke nodes and dual worlds that serve as complementary counterparts. Each regular world w is associated with a dual world d(w), and the model features two independent relations: a forward accessibility R governing the evolution of truth in regular worlds, and a backward (or “dual”) relation R⁻ governing the propagation of falsity in dual worlds.

The central technical device is the notion of “forcing” and “anti‑forcing.” In a regular world, a formula φ is forced (w ⊨ φ) exactly as in the intuitionistic case: every R‑successor must also force φ. In a dual world, anti‑forcing (u ⊭ φ) is defined dually: every R⁻‑successor must also reject φ. Because each world has a unique dual, the model guarantees that for any formula φ, either φ is forced or its negation is forced at a given world, thereby validating the law of excluded middle at the semantic level. This duality also ensures that double‑negation elimination holds automatically: if ¬¬φ is forced, then φ must be forced, since the dual of a dual world returns to the original regular world.

Building on this semantics, the authors develop a proof system that mirrors natural deduction but is enriched with two families of inference rules: forcing rules that operate within regular worlds and anti‑forcing rules that operate within dual worlds. Crucially, the system does not require the cut rule; the authors prove a cut‑free completeness theorem by constructing, for any non‑derivable sequent, a counter‑model consisting of a regular world together with its dual that falsifies the sequent. The proof proceeds by a constructive, inductive construction of a saturated set of formulas, then interpreting this set as the truth assignment in a regular world, while the complementary set defines the anti‑forcing behavior in the associated dual world. This method yields a fully constructive completeness proof, unlike classical Hilbert‑style arguments that often rely on non‑constructive maximal consistent sets.

The paper also discusses several potential applications. In automated theorem proving, the dual‑world architecture allows a search algorithm to explore forcing and anti‑forcing branches in parallel, effectively halving the search space for classical tautologies. In program verification, specifications can be interpreted as forced properties in regular worlds, while error conditions correspond to anti‑forced properties in dual worlds, enabling a symmetric treatment of safety and liveness. The authors further speculate on extensions to modal and quantum logics, noting that the regular/dual dichotomy resembles the superposition‑measurement duality in quantum theory, and that the forcing/anti‑forcing mechanism could model probabilistic collapse without abandoning a Kripke‑style relational semantics.

In conclusion, the work provides a clean, constructive semantic foundation for classical logic that preserves the intuitive relational picture of Kripke models while overcoming their intuitionistic limitations. The cut‑free completeness result, together with the explicit dual‑world construction, opens new avenues for both theoretical investigations—such as exploring richer logical connectives or categorical interpretations—and practical implementations in proof assistants and verification tools. Future research directions include optimizing the dual‑world search strategies, integrating the framework with existing proof assistants, and extending the semantics to handle higher‑order or dependent types.