Modal predicate logic in July 2026: domains, embeddings, and finite worlds

The short version

Three 2026 preprints make the same point from different directions: quantified modal logic becomes manageable only when its domain assumptions are made explicit. One mechanizes first-order modal logic in Isabelle/HOL, one builds cut-free proof systems that track changing inner domains, and one transfers finite quantified Kripke models into arithmetic. Together they turn the Barcan debate into a set of semantic and proof-theoretic choices rather than a slogan.

The entries below are arXiv preprints, and the dates are the posting dates visible on their arXiv records. The selection is organized by the formal problem each paper solves, not by citation count or an attempt at an exhaustive bibliography.

The Romeo–Juliet test, stated carefully

Let

L (x, y)

mean that

x

loves

y

, with

r

and

j

naming Romeo and Juliet. The tempting pair is:

□ L (r, j) and ◊ \neg\exists y L (r, y) .

It is a useful stress test, but the two formulas are not automatically compatible in every quantified modal language. Under constant-domain, rigid-designator semantics,

j

remains in the domain at every accessible world. If

□ L (r, j)

is true, then every accessible world has at least one loved object, namely

j

; therefore

◊ \neg\exists y L (r, y)

is false.

To express the intended contingentist intuition, we need to say what it means for an object to exist at a world. A common two-domain setup has an outer domain

U

of possible objects and a world-relative inner domain

D_{w} \subseteq U

of existing objects. The existence predicate can be written as:

E t := \exists x (x = t) .

A guarded version of the intuition is then:

□ (E r \land E j \to L (r, j))

alongside:

◊ (E r \land \neg E j \land \neg\exists y (E y \land L (r, y))) .

The first formula says that whenever both people exist, Romeo's love for Juliet is necessary. The second says that there is an accessible world in which Romeo exists, Juliet does not, and Romeo loves nobody who exists there. This is not a free pass for the original unguarded

□ L (r, j)

: it makes the semantic choice visible. The 2026 work below is largely about exactly this choice.

The formal hinge: domains and Barcan formulas

A varying-domain Kripke model can be written as

M = ⟨ W, R, {D_{w}}_{w \in W}, ⊩ ⟩

, where

W

is a set of worlds,

R

is accessibility, and

D_{w}

is the inner domain at

w

. The satisfaction clauses used in the recent proof-theoretic work are:

M, w, σ ⊩ \exists x φ iff \exists o \in D_{w} M, w, σ^{x ▹ o} ⊩ φ,

M, w, σ ⊩ \forall x φ iff \forall o \in D_{w} M, w, σ^{x ▹ o} ⊩ φ,

M, w, σ ⊩ □ φ iff \forall v (w R v \to M, v, σ ⊩ φ) .

The assignment stays fixed when the world shifts, while the range of quantification changes with

D_{w}

. That separation is where the Barcan formulas get their force:

BF : \forall x □ φ (x) \to □ \forall x φ (x),

CBF : □ \forall x φ (x) \to \forall x □ φ (x) .

The domain condition determines which direction is sound:

Domain condition along $w R v$	Formula supported
Increasing: $D_{w} \subseteq D_{v}$	Converse Barcan, $CBF$
Decreasing: $D_{v} \subseteq D_{w}$	Barcan, $BF$
Constant: $D_{w} = D_{v}$	Both $BF$ and $CBF$
No inclusion condition	Neither direction is generally forced

This is the cleanest formal bridge to the necessitism/contingentism dispute. A metaphysical argument may favor one schema, reject it, or accept both, but the logic must then say what happens to domains, assignments, names, and existence predicates. The papers do not settle the metaphysics; they expose the commitments that a metaphysical reading places on the proof system.

Christoph Benzmüller and Daniel Kirchner, posted 12 July 2026. Read the extended preprint.

What the paper does

The authors extend a previous deep-and-shallow embedding methodology from propositional modal logic to first-order modal logic in Isabelle/HOL. The formal development works with constant-domain Kripke semantics and places three representations side by side:

a deep embedding, where formulas are an inductive datatype;
a heavyweight maximal-shallow embedding, carrying the model parameters explicitly; and
a lightweight minimal-shallow embedding, packaged as an Isabelle/HOL locale.

The minimal locale is designed so that quantifying over all its interpretations recovers exactly deep validity. That is the paper's central faithfulness claim, not merely a translation into a convenient host language.

The mechanism kept visible

The object-language grammar is:

φ, ψ ::= r (x, y) ∣ \neg φ ∣ (φ \land ψ) ∣ \exists x . φ ∣ □ φ .

The other familiar operators are defined by duality. In a constant-domain model

⟨ W, R, I, D ⟩

, existential quantification and necessity are interpreted as:

⟨ W, R, I, D ⟩, g, w ⊨ \exists x . φ iff \exists d \in D ⟨ W, R, I, D ⟩, g [x \leftarrow d], w ⊨ φ,

⟨ W, R, I, D ⟩, g, w ⊨ □ φ iff \forall v \in W (R (w, v) \to ⟨ W, R, I, D ⟩, g, v ⊨ φ) .

Because

D

is shared by every world and

g : V \to D

is world-independent, both Barcan directions are valid in the setting mechanized by the paper. The development also formalizes the parts that are easy to hide in an informal account: free, bound, and fresh variables; capture-avoiding substitution; alphabetic renaming; the substitution lemma; and size-based induction principles.

The technically unusual step is a mechanized countable downward Löwenheim–Skolem theorem. It supplies a countable elementary subdomain

N \subseteq D

preserving truth for the relevant formulas. That solves a surjectivity problem: the variable set is represented by

N

, so a variable assignment cannot be onto an uncountable individual domain. The paper uses the elementary subdomain to lift faithfulness back to the full domain.

What is demonstrated, and what is not

The Isabelle experiments confirm the K axiom, necessitation, the Barcan and converse Barcan formulas, and frame correspondences for reflexive, transitive, and symmetric frames. The scope is deliberately narrower than general quantified modal logic: the semantics has constant domains and rigid assignments, and the varying-domain extension is left for future work. The Löwenheim–Skolem construction also assumes a countable world domain in the present development.

Read this first if: you want a machine-checked account of how syntax, substitution, Kripke semantics, and HOL automation fit together.

2. Making changing domains proof-theoretic

Tim S. Lyon and Eugenio Orlandelli, posted 20 April 2026. Read the journal-version preprint.

What the paper does

This paper develops cut-free nested sequent systems for a broad class of quantified modal logics with equality. Its models distinguish an outer domain

U

of possible objects from inner domains

D_{w}

of objects that exist at each world. The framework covers increasing, decreasing, constant, empty, and non-empty domain conditions, together with seriality and generalized path conditions.

The proof systems add signatures, which are multisets of terms, to nested sequents. They also use reachability rules parameterized by formal grammars, allowing formulas and terms to be propagated or searched for along paths in a nested sequent. The result is a uniform system rather than a separate ad hoc calculus for every domain condition.

The Barcan result is structural

The paper's frame-to-axiom table makes the domain issue explicit:

ID : \forall w, v (w R v \to D_{w} \subseteq D_{v}) ⟺ □ \forall x φ \to \forall x □ φ,

DD : \forall w, v (w R v \to D_{w} \supseteq D_{v}) ⟺ \forall x □ φ \to □ \forall x φ .

Adding both inclusion conditions yields constant inner domains. The system also contains the extended Barcan rule. The authors show that the standard universal-quantifier rule subsumes it, so the nested systems naturally capture logics with constant outer domains. They establish soundness and completeness for the target class, invertibility of all rules, height-preserving admissibility results, and a syntactic cut-elimination theorem.

This is important for the Romeo–Juliet test. If

j

is absent from an accessible inner domain, then a formula quantified over existing objects cannot silently treat

j

as an ordinary witness. The signature and existence machinery force the proof to account for which terms are available at which worlds.

Limits and reading decision

The paper's generality is semantic rather than metaphysical. It gives proof systems for a wide class of domain conditions, but it does not choose between necessitism and contingentism. Its constant outer domain is part of the setting forced by the extended Barcan rule used in the calculus; inner domains can still vary.

Read this first if: your main question is how a proof system can distinguish

BF

CBF

, and constant-domain reasoning without collapsing them into one calculus.

3. Finite quantified worlds inside arithmetic

Haruka Kogure and Taishi Kurahashi, posted 28 April 2026. Read the preprint.

What the paper does

The authors study arithmetical completeness for finite Kripke models of quantified modal logic. The motivation comes from provability logic: replace the modal operator with an arithmetic provability predicate and ask how much of a quantified Kripke model can be represented inside arithmetic.

For each conversely well-founded finite quantified Kripke model, they construct a

Σ_{2}

Fefermanian provability predicate and an arithmetical interpretation embedding the model into arithmetic. For finite constant-domain Kripke models, they construct a

Σ_{1}

provability predicate satisfying the global derivability condition

D 2^{G}

and an arithmetical interpretation with the same embedding role.

The formal bridge

Their quantified modal language includes first-order quantifiers and a primitive

□

, with:

◊ A := \neg □ \neg A .

A Kripke frame is

F = (W, R, {D_{w}}_{w \in W})

, with increasing domains along accessibility:

x R y \to D_{x} \subseteq D_{y} .

The constant-domain case is the special situation

D_{x} = D_{y}

for all worlds. The base quantified modal logic

QK

contains classical first-order logic, the K axiom

□ (A \to B) \to (□ A \to □ B),

modus ponens, generalization, and necessitation. Its constant-domain extension adds the Barcan schema

\forall x □ A (x) \to □ \forall x A (x) .

The arithmetic interpretation

f

preserves Boolean connectives and quantifiers. The modal clause is the decisive one:

f (□ A (x_{0}, \dots, x_{n - 1})) := Pr_{T} (┌ f (A) (\overset{x}{˙}_{0}, \dots, \overset{x}{˙}_{n - 1}) ┐),

where

Pr_{T}

is a provability predicate and dotted variables are substituted by numerals in the Gödel code. The authors identify the quantifier and numeral-substitution cases as the main obstacle that must be overcome beyond the propositional setting.

Boundary of the result

The results concern finite models and specific classes of provability predicates. The paper records an open problem about removing the conversely well-foundedness assumption from its

Σ_{2}

construction. It also recalls that the full quantified provability logic of Peano Arithmetic is more complicated than simply taking a familiar quantified provability logic and declaring the two identical.

Read this first if: you care about the connection between quantified Kripke semantics, Gödel coding, and provability interpretations rather than metaphysical modality alone.

What this changes for the philosophical debate

The current results do not replace the necessitism/contingentism debate with a technical theorem. They make its hidden decisions harder to ignore.

A constant-domain system validates both $BF$ and $CBF$ , but that is a semantic consequence of $D_{w} = D_{v}$ , not an argument that every object must exist necessarily.
An increasing-domain system supports $CBF$ , while a decreasing-domain system supports $BF$ . Rejecting one schema therefore requires saying which domain inclusions are being rejected, or which quantifier semantics is being used instead.
If names are rigid across worlds, $□ L (r, j)$ treats $j$ differently from a merely possible object. If names are world-relative, or if existence is represented by an inner-domain predicate, the formula must be guarded or evaluated in a free-logic setting.
The proof-theoretic papers show that these are not cosmetic choices. They change which quantifier rules are sound, which substitutions are admissible, and which countermodels can be extracted.

That is the live research signal in this batch: the philosophical question is being pressed through semantics, proof theory, and mechanization at once. The next useful comparison is not simply “BF or CBF?” It is “BF or CBF under which domain, assignment, existence, and proof-system assumptions?”

A reading order

Start with Benzmüller and Kirchner for a machine-checked constant-domain baseline.
Move to Lyon and Orlandelli for varying inner domains, equality, reachability rules, and cut elimination.
Read Kogure and Kurahashi if the provability-logic and arithmetic side is part of your project.

The compact lesson is formal rather than rhetorical: before deciding whether Romeo and Juliet's love is necessary, specify what counts as an object at each world and what your quantifiers are allowed to range over.

Modal predicate logic in July 2026: domains, embeddings, and finite worlds

The short version

The Romeo–Juliet test, stated carefully

The formal hinge: domains and Barcan formulas

What the paper does

The mechanism kept visible

What is demonstrated, and what is not

2. Making changing domains proof-theoretic

What the paper does

The Barcan result is structural

Limits and reading decision

3. Finite quantified worlds inside arithmetic

What the paper does

The formal bridge

Boundary of the result

What this changes for the philosophical debate

A reading order

Fuentes de referencia

Contenido relacionado

The short version

The Romeo–Juliet test, stated carefully

The formal hinge: domains and Barcan formulas

1. Mechanizing first-order modal logic in HOL

What the paper does

The mechanism kept visible

What is demonstrated, and what is not

2. Making changing domains proof-theoretic

What the paper does

The Barcan result is structural

Limits and reading decision

3. Finite quantified worlds inside arithmetic

What the paper does

The formal bridge

Boundary of the result

What this changes for the philosophical debate

A reading order

Fuentes de referencia

Contenido relacionado