Duality for Constructive Modal Logics: from Sahqlvist to Goldblatt-Thomason

Jim de Groot, Ian Shillito, Ranald Clouston, [1em] Mathematical Institute, University of Bern, Bern, Switzerland[-.2em] [email protected], [.2em] School of Computer Science, University of Birimingham, Birmingham, UK[-.2em] [email protected], [.2em] School of Computing, The Australian Nati · 2026

Read Duality for Constructive Modal Logics: from Sahqlvist to Goldblatt-Thomason by Jim de Groot, Ian Shillito, Ranald Clouston, [1em] Mathematical Institute, University of Bern, Bern, Switzerland[-.2em] [email protected], [.2em] School of Computer Science, University of Birimingham, Birmingham, UK[-.2em] [email protected], [.2em] School of Computing, The Australian Nati (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-03762.txt.

Introduction

Jim de Groot, Ian Shillito, and Ranald Clouston study the constructive modal logic CK, a logic of possibility and necessity built on constructive rather than classical reasoning. Their approach is semantic: they build a categorical duality, a systematic two-way bridge between the logic's algebraic side and its birelational models, which use two relations to interpret proofs and possibilities.

That bridge then does concrete work. The authors prove Sahlqvist-style correspondence results, which match logical formulas to properties of frames, together with completeness results. They also prove a Goldblatt-Thomason style theorem about which classes of frames the logic can define. Duality theorems are prized in logic because one construction can transport many results at once; here it organizes the theory of a logic that constructive mathematics and computer science both care about.

Catalog record

CREATORJim de Groot, Ian Shillito, Ranald Clouston, [1em] Mathematical Institute, University of Bern, Bern, Switzerland[-.2em] [email protected], [.2em] School of Computer Science, University of Birimingham, Birmingham, UK[-.2em] [email protected], [.2em] School of Computing, The Australian Nati
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE148 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-03762

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Duality for Constructive Modal Logics: from Sahqlvist to Goldblatt-Thomason Jim de Groot, Ian Shillito, Ranald Clouston, [1em] Mathematical Institute, University of Bern, Bern, Switzerland[-.2em] [email protected], [.2em] School of Computer Science, University of Birimingham, Birmingham, UK[-.2em] [email protected], [.2em] School of Computing, The Australian National University, Canberra, Australia[-.2em] [email protected] Abstract We carry out a semantic study of the constructive modal logic 𝖢𝖪 \mathsf{CK} . We provide a categorical duality linking the algebraic and birelational semantics of the logic. We then use this to prove Sahlqvist style correspondence and completeness results, as well as a Goldblatt-Thomason style theorem on definability of classes of frames. 1 Introduction The question of how best to define a basic intuitionistic version of modal logic, particularly in the presence of the sometimes omitted \Diamond (possibility) operator, has received a wide range of differing answers [ 16 , 48 , 3 , 32 , 2 ] . A plausible minimal answer is Constructive \mathsf{K} ( \mathsf{CK} ) [ 3 ] , which can be defined by extending the usual axioms of intuitionsitic propositional logic with two further axioms ( \mathsf{K_{\Box}} ) \Box(\varphi\to\psi)\to(\Box\varphi\to\Box\psi) ( \mathsf{K_{\mathchoice{\raisebox{-0.9pt}{$\displaystyle\Diamond$}}{\raisebox{-0.9pt}{$\Diamond$}}{\raisebox{-0.5pt}{$\scriptstyle\Diamond$}}{\raisebox{-0.2pt}{$\scriptscriptstyle\Diamond$}}}} ) \Box(\varphi\to\psi)\to(\mathchoice{\raisebox{-0.9pt}{$\displaystyle\Diamond$}}{\raisebox{-0.9pt}{$\Diamond$}}{\raisebox{-0.5pt}{$\scriptstyle\Diamond$}}{\raisebox{-0.2pt}{$\scriptscriptstyle\Diamond$}}\varphi\to\mathchoice{\raisebox{-0.9pt}{$\displaystyle\Diamond$}}{\raisebox{-0.9pt}{$\Diamond$}}{\raisebox{-0.5pt}{$\scriptstyle\Diamond$}}{\raisebox{-0.2pt}{$\scriptscriptstyle\Diamond$}}\psi) where \Box is the necessity operator, and by adding Modus Ponens and the usual necessitation rule of modal logic (if \varphi is a theorem, then so is \Box\varphi ). Alternatively, \mathsf{CK} arises proof-theoretically from restricting a standard sequent calculus for the classical modal logic \mathsf{K} to single conclusions. Most competing notions of basic intuitionistic modal logic can be expressed as axiomatic extensions of \mathsf{CK} , but \mathsf{CK} is not too minimal to be given birelational (i.e. Kripke style) semantics [ 35 ] . 1 1 1 A non-example of this is the logic of Božić and Došen [ 8 ] , which is incomparable to \mathsf{CK} but has only been given trirelational semantics in which \Box and \Diamond do not talk about the same relation. It was recently shown by the authors of this paper [ 29 ] how to capture the most well known alternative notions of basic intuitionistic modal logic not just as axiomatic extensions of \mathsf{CK} , but by corresponding conditions on the birelational frames for \mathsf{CK} , creating a new semantic understanding of this zoo of competing logics. The logic \mathsf{CK} and its extensions have applications ranging from knowledge representation [ 35 ] to various flavours of constructive epistemic logic [ 50 , 1 , 37 ] to modelling parallel computation [ 48 , 49 ] and evaluation [ 39 ] . Besides, the \Diamond -free fragment of \mathsf{CK} , which can be axiomatised simply by dropping the K ◇ axiom, has been studied extensively. By contrast, the \Diamond -free fragment of the alternative basic intuitionistic modal logic Intuitionistic \mathsf{K} [ 16 ] is somewhat mysterious, with no known finite axiomatisation [ 24 , 12 ] . In this paper, we continue the semantic study of \mathsf{CK} : • In Section 2 we introduce the main constructions of our algebraic and frame semantics. • In Section 3 we derive a categorical duality between the algebraic semantics of \mathsf{CK} and a suitable notion of descriptive \mathsf{CK} -frames. We then use this duality to further study \mathsf{CK} : • In Section 4 we prove Sahlqvist style correspondence and completeness results for the logic. The latter theorem relies on the general completeness theorem of extensions of \mathsf{CK} with respect to a suitable class of descriptive \mathsf{CK} -frames, which follows immediately from the duality. • In Section 5 we prove a Goldblatt-Thomason style theorem, which describes when certain classes of \mathsf{CK} -frames are definable by formulas. This time we rely on the duality to transfer Birkhoff’s variety theorem from algebras to \mathsf{CK} -frames. Duality In the field of (modal) logic, duality theorems are used to bridge the gap between the algebraic and geometric (frame-based) views of a logic. They can be used to prove completeness results, as well as (analogues of) Sahlqvist completeness theorems, bisimilarity-somewhere-else, and Goldblatt-Thomason style theorems. The study of dualities dates back to Stone’s representation theorem for Boolean algebras [ 46 ] , which establishes a categorical duality between Boolean algebras and certain topological spaces (now called Stone spaces). Subsequent notable results include the McKinsey-Tarski representation theorem for closure algebras [ 34 ] , the Priestley duality theorem [ 40 ] linking distributive lattices and spaces now known as Priestley spaces, and Esakia duality for Heyting algebras (the algebraic semantics of intuitionistic logic) [ 14 , 15 ] . Dualities for modal logics often build on a duality for the underlying propositional logic. The first such result was Goldblatt’s duality for modal algebras [ 20 , 21 ] , which extends Stone duality. This establishes a duality between the algebraic semantics of normal modal logic on the one hand, and Kripke frames with extra structure on the other. The resulting frames are called descriptive Kripke frames, and consist of a Kripke frame together with a collection of so-called admissible subsets of the frame that satisfy certain conditions. Similar dualities have been derived for monotone modal logic [ 31 ] and neighbourhood logic [ 13 ] , where in each case the geometric side of the duality is given by frames for the logic with a suitable collection of admissible sets. Mirroring this, various dualities for modal extensions of positive logic were derived by extending Priestley duality [ 9 , 25 , 28 ] . When working over intuitionistic logic, one often builds on Esakia duality [ 38 , 52 , 26 ] . Again, the geometric side of these dualities is given by frames of the logic with a collection of admissible subsets. But in the case of \mathsf{CK} something interesting happens: the duality between \mathsf{CK} -algebras and a descriptive analogue of \mathsf{CK} -frames cannot piggy-back on Esakia duality. The intuitive reason is that the birelational semantics for \mathsf{CK} makes crucial use of the fact that the intuitionistic accessibility relation is a preorder, while Esakia duality requires a partial order. More precisely, the \mathsf{CK} -frame dual to a \mathsf{CK} -algebra is not based on a set of prime filters, but on a set of segments . These are pairs (\mathfrak{p},\Gamma) consisting of a prime filter \mathfrak{p} together with a set of prime filters \Gamma . The intuitionistic accessibility relation is then given by inclusion of the first argument, and since we can have different segments headed by the same prime filter, this gives rise to a preorder but not a partial order. Hence the duality of this paper is not merely a mathematical tool for further study of the logic, but also has interesting features in its own right. 2 2 2 It would still be possible to piggy-back on Esakia if we used a different semantics, for example trirelational frames. Since we insist on using the birelational \mathsf{CK} -frames we disregard such alternatives. Sahlqvist correspondence and completeness After the introduction of the Kripke semantics for classical modal logic, it was rapidly noticed that some modal axioms, such as \Box p\to p and \Box p\to\Box\Box p , correspond to properties on frames, e.g. reflexivity and transitivity. These insights led the community to wonder about the existence of a general theory of correspondence. In 1975, Sahlqvist gathered the results from his

Read the complete text