🔢 mathematics

Some prospects for semiproducts and products of modal logics

This paper presents new examples and counterexamples regarding the axiomatization and finite model property of products and semiproducts of propositional modal logics with S5, utilizing local tabularity and bisimulation games to establish decidability results for specific fragments of predicate modal logics.

Original authors: Valentin Shehtman, Dmitry Shkatov

Published 2026-07-21

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Valentin Shehtman, Dmitry Shkatov

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are trying to build a massive, perfect Lego city. In the world of computer science and mathematics, there is a special branch called "modal logic" that acts like the instruction manual for how things can be possible or necessary. Think of it as the rulebook for a game where you don't just say "this is true," but "this is true in every possible world." Now, imagine you want to combine two different rulebooks: one that describes a world where everything is connected in a specific way, and another that describes a world where everything is connected to everything else (like a universal "all-knowing" perspective).

This paper dives into the tricky business of merging these two rulebooks. The authors are asking a very specific question: When we smash these two logical systems together, do we get a new, clean system that we can easily understand and solve? Or does the combination create a chaotic mess that breaks the rules? This matters because these logical systems are the hidden engines behind how we verify computer software and understand the structure of language. If the combined system is "well-behaved," we can write programs to check if our logic is sound. If it's messy, we might get stuck in an infinite loop, never knowing if our answer is right or wrong. The authors are essentially testing the structural integrity of these logical "Lego cities" to see which combinations hold up and which crumble.

The Great Logic Mash-Up: When Worlds Collide

In this paper, two mathematicians, Valentin Shehtman and Dmitry Shkatov, act like master architects testing the stability of new logical structures. They are mixing a specific type of logic (let's call it "Logic A") with a very powerful, all-encompassing logic called S5. Think of S5 as a "Universal Remote Control" for logic; it represents a world where every possibility is reachable from every other point, like a room where you can instantly teleport to any other spot.

The authors are investigating two ways to mix these logics:

The Product: A perfect, grid-like combination where the rules of both worlds apply strictly side-by-side.
The Semiproduct: A slightly looser, more flexible combination where the rules interact but might not be perfectly symmetrical.

Their goal is to find out if these new, mixed logics are "axiomatizable in the minimal way." In plain English, this means: Can we write down a short, simple list of rules that perfectly describes the new system without needing an infinite number of instructions? If we can, the system is "decidable," meaning a computer can eventually solve any problem posed to it. If not, the system might be a nightmare that no computer can ever fully solve.

The Good News: Building Stable Towers

The authors discovered that for certain types of "Logic A," the mash-up works beautifully. Specifically, if "Logic A" has a "finite depth" (imagine a tree that can only grow so tall before it stops), the resulting mixed logic is stable.

They used a clever technique involving "bisimulation games" to prove this. Picture this as a game of "spot the difference" played between two detectives. If the detectives can't find any difference between two logical worlds after a certain number of moves, the worlds are effectively the same. The authors showed that for these finite-depth logics, the game always ends quickly. This proves that the new mixed logics have the Finite Model Property (FMP).

What does FMP mean for a teenager? It means that to test if a statement is true in this new system, you don't need to check an infinite universe. You only need to check a tiny, finite model. It's like proving a bridge is safe by testing a small, perfect scale model rather than building the whole thing first. Because of this, the authors confirmed that for these specific logics, we can definitely write a computer program to decide if any statement is true or false. They also found that this works for a specific family of logics involving a rule called Ath (which sounds like a rule about how paths connect), showing that even with these extra rules, the system remains stable and solvable.

The Bad News: The Crumbling Foundations

However, the story isn't all happy endings. The authors also found some "counterexamples"—combinations that simply don't work. They proved that if you take certain other logics (specifically those that sit between two complex rules called □T and SL4) and mix them with S5, the result is a disaster.

In these cases, the "minimal" list of rules fails. The mixed logic becomes too complex to be described simply, and it loses the nice property of being "semiproduct-matching." The authors showed that even though these individual logics are well-behaved on their own, when you try to combine them with the "Universal Remote" (S5), they break the rules. It's like trying to mix oil and water; no matter how hard you stir, they refuse to form a single, stable mixture.

One of the most surprising findings is that even logics that are "Horn axiomatizable" (a fancy way of saying they follow a very specific, simple type of rule) can fail when mixed with S5. This rules out a hopeful idea that all simple logics would play nicely together. The authors explicitly showed that for logics like K + Altn (where n is 3 or more), the combination is neither product-matching nor semiproduct-matching. The resulting structure is too messy to be captured by a simple set of rules.

The Takeaway: A Map of What Works and What Doesn't

So, what is the final verdict? Shehtman and Shkatov have drawn a new map of the logical landscape. They have identified a safe zone where mixing logics creates a stable, solvable system that computers can handle, provided the original logic isn't too deep or complex. They proved that for these safe zones, the "1-variable fragments" (simplified versions of the logic) are also solvable.

But they also marked the danger zones. They showed that there are infinite families of logics that, when mixed with S5, create systems that cannot be described simply. They didn't just guess this; they provided rigorous mathematical proofs using games and frame constructions to demonstrate exactly where the logic breaks.

In the end, this paper doesn't solve every problem in the universe of logic, but it gives us a very clear guide on which combinations are worth building and which ones are destined to collapse. It tells us that while we can build some magnificent logical towers by mixing these systems, we must be careful not to mix the wrong ingredients, or the whole structure might fall apart. For anyone trying to verify software or understand the deep structure of reasoning, this map is an essential tool for knowing where it's safe to step.

Technical Summary: Some prospects for semiproducts and products of modal logics

Problem Statement
The paper investigates the properties of combined modal logics, specifically focusing on the products ( $\Lambda \times S5$ ) and semiproducts ( $\Lambda \rhd S5$ ) of a propositional modal logic $\Lambda$ with the logic S5. The central problem addresses the "matching" conditions: determining when the syntactic combination of logics (via fusion and specific interaction axioms) coincides with the semantic combination (the logic of the class of product or semiproduct frames).

Specifically, the authors examine whether $\Lambda$ and $S5$ are product-matching (where $[\Lambda, S5] = \Lambda \times S5$ ) or semiproduct-matching (where $\Lambda | S5 = \Lambda \rhd S5$ ). A critical motivation for this study is the connection to 1-variable fragments of predicate modal logics ( $Q\Lambda$ and $Q\Lambda + \text{Barcan}$ ). If $\Lambda$ and $S5$ are matching, the 1-variable fragment of the corresponding predicate logic is axiomatized in the "minimal" way and inherits the Finite Model Property (FMP) and decidability from the propositional (semi)product.

Methodology
The authors employ a combination of frame-theoretic constructions, algebraic logic, and model-theoretic techniques:

Frame Products and Semiproducts: The analysis relies on the definition of product frames ( $F_1 \times F_2$ ) and semiproducts (subframes of products where the horizontal relation is preserved).
Bisimulation Games: To establish local tabularity for logics of finite depth, the authors utilize bisimulation games (referencing techniques from [7]). This is a crucial step in proving the Finite Model Property (FMP).
p-morphic Images: The proofs for semiproduct-matching often involve demonstrating that specific frames (such as countable rooted frames satisfying certain axioms) are p-morphic images of semiproducts of trees and clusters.
Counterexample Construction: To disprove matching properties, the authors construct specific frames (e.g., modified trees and clusters) that validate the semantic logic but fail to validate the syntactic logic, often utilizing Fine-Jankov formulas.

Key Contributions and Results

New Positive Examples (Semiproduct-Matching):
The paper identifies new classes of logics that are semiproduct-matching with $S5$ and possess the product/semiproduct FMP.
- Finite Depth Logics: It is shown that if $\Lambda$ is of finite depth (i.e., $\vdash \Box^n \bot$ ), then the logics $\Lambda \rhd S5$ and $\Lambda | S5$ enjoy the FMP. This is derived from the local tabularity of $(K + \Box^n \bot) | S5$ .
- Thick Frames (Ath): The authors introduce the axiom $Ath := \Diamond\Diamond p \to \Box\Diamond p$ , defining "thick" frames. They prove that $K + Ath$ and $K + Ath + \Box^n \bot$ are semiproduct-matching with $S5$ .
- Implication: These results imply that the 1-variable fragments of the corresponding predicate logics ( $Q\Lambda$ and $Q\Lambda + \text{Barcan}$ ) are decidable and possess the FMP.
New Counterexamples (Non-Matching):
The paper presents two infinite families of logics that are not semiproduct-matching with $S5$ , challenging the assumption that Horn axiomatizability guarantees matching.
- The $\Box T$ to $SL4$ Range: For any logic $\Lambda$ such that $\Box T \subseteq \Lambda \subseteq SL4$ (where $\Box T = K + \Box(\Box p \to p)$ and $SL4 = K4 + \Diamond p \leftrightarrow \Box p$ ), $\Lambda$ and $S5$ are not semiproduct-matching. This includes Horn axiomatizable logics like $\Box T$ , $K5$ , and $K45$ .
- Alten and Box-Depth Logics: For $\Lambda$ such that $K + Alt_n \subseteq \Lambda \subseteq K + Alt_n + \Box^m \bot$ (with $n \ge 3, m \ge 2$ ), $\Lambda$ and $S5$ are neither product- nor semiproduct-matching.
- Significance of Counterexamples: These results demonstrate that even complete, Horn axiomatizable logics may fail to be semiproduct-matching, providing a negative answer to a potential analogue of the "product-matching" theorem (Theorem 2.8) for semiproducts.
Finite Model Property (FMP) for (Semi)Products:
Theorem 6.1 establishes that for $\Lambda = K + Ath$ and $\Lambda = K + \Box^n \bot + Ath$ , the (semi)product of $\Lambda$ with $S5$ has the (semi)product FMP. The proof involves showing that finite rooted frames for these logics are p-morphic images of "straight" frames isomorphic to semiproducts of $\Lambda$ -frames and clusters.

Significance and Claims
The paper claims to advance the systematic investigation of combined modal logics by:

Clarifying the limits of Horn axiomatizability: While Horn axiomatizable logics are known to be product-matching (Theorem 2.8), this work shows they are not necessarily semiproduct-matching.
Establishing decidability: By proving the FMP for specific (semi)products, the authors secure the decidability of the 1-variable fragments of the associated predicate modal logics.
Local Tabularity: The use of bisimulation games to prove local tabularity for finite-depth logics provides a robust method for establishing the FMP in this context.

The authors modestly suggest that their main results regarding semiproduct-matching and FMP might be transferable to extensions of the provability logic GL (specifically mentioning the Artemov–Japaridze logic and Solovay's logic), though this is presented as a conjecture rather than a proven result within the paper.

The Great Logic Mash-Up: When Worlds Collide

The Good News: Building Stable Towers

The Bad News: The Crumbling Foundations

The Takeaway: A Map of What Works and What Doesn't

More like this