The paper tackles a gap in our understanding of non-contingency logic, a system of reasoning that replaces the standard "necessarily true" operator of modal logic with one that asks whether a statement is either necessarily true or necessarily false. This kind of logic turns out to be surprisingly useful in several areas: in reasoning about knowledge, the operator captures the idea of "knowing whether" something is the case (as opposed to knowing that it is true); in the study of mathematical provability, it connects to questions about decidability. While researchers have spent considerable effort understanding what these logics mean and what they can express, the question of how to formally prove things within them has received far less attention.
The authors address this by building a systematic, unified framework for constructing formal proof systems for a wide family of non-contingency logics. The key technical tool they introduce is called generalized path conditions, a way of describing structural properties of the underlying logical frameworks using a kind of grammar. Many common assumptions one might make about how logical worlds or states relate to each other can be captured this way. For any chosen collection of these conditions, the authors can automatically generate a corresponding proof system, called a labelled sequent calculus, which is a formal procedure for deriving conclusions from premises using explicit labels to track the underlying structure.
What makes the contribution notable is its uniformity. All the proof systems the authors construct share the same core logical rules and differ only in a single additional structural rule that reflects the particular assumptions being made about the logical framework. The authors prove that each such system is both sound (it only proves things that are actually true under those assumptions) and complete (it can prove everything that is true). This gives researchers a reliable, modular toolkit for working with non-contingency logic across many different settings, potentially opening the door to practical applications in areas like automated reasoning about knowledge and decidability.