Infinitary Cut Elimination

Updated 8 February 2026

Infinitary Cut Elimination is a suite of proof-theoretic techniques that remove the cut rule using transfinite induction in non-wellfounded and ω-rule-extended sequent calculi.
It guarantees strong normalization and the subformula property, which are crucial for fixed-point logics, transparent truth predicates, and cyclic proof systems.
The methodology supports ordinal analysis and proof compression, impacting applications in automated reasoning and formal verification.

Infinitary cut elimination is a suite of proof-theoretic techniques and results establishing the admissibility or eliminability of the cut rule in infinitary, non-wellfounded, or ω-rule-extended sequent calculi. Such systems play a central role in the proof theory of logics with fixed-point operators, transparent truth predicates, infinitary conjunction/disjunction, cyclic proofs, or generalized forms of induction and coinduction. Infinitary cut-elimination establishes both strong normalization and the subformula property in highly expressive proof systems, including those featuring nontermination or infinite proof trees—often as a prerequisite for semantics via minimal fixed points, Herbrand systems, or ordinal analyses.

1. Foundational Concepts and Typical Settings

Infinitary proof systems arise in various settings: classical and intuitionistic first-order logics with ω-rules, fixed-point logics (µ-calculi), logics of truth with predicates, and linear logics with infinite derivations. Such systems permit either infinitely branching rules (e.g., ω-rules for quantifiers or fixed points) or allow infinite-depth proofs as long as certain global well-behavedness conditions are met, such as progressivity along infinite branches or compliance with a trace condition (Nicolai, 2020, Curzi et al., 1 Feb 2026, Afshari et al., 13 Oct 2025, Baelde et al., 2020).

Sequents and inferences are often enriched, for example by term-encoding sentences as numerals (as in transparent truth systems), or by explicit treatment of addresses/formula occurrences (as in μ-calculus systems). Various forms of initial sequents and structural rules (such as contraction, weakening, or their restriction) crucially impact the behavior and analyzability of infinitary cut elimination (Nicolai, 2020, Afshari et al., 13 Oct 2025).

A key motivation is that, due to the presence of fixed points or infinitary constructs, cut is not trivially eliminable with the strictly finitary syntactic techniques of standard Gentzen-style calculi.

2. Infinitary Sequent Calculi and ω-rules

Infinitary cut elimination is justified in sequent calculi that extend classical rules with infinitary inferences. For example, in systems of transparent truth, one moves from a finitary calculus with ordinary logical and truth rules to an infinitary calculus incorporating ω-rules for universal quantification, and arithmetically rich initial sequents that serve as atomic axioms indexed by numerals (Nicolai, 2020). The ω-rule in these systems typically states:

$\frac{\forall t \in \text{closed terms}: \;\; \Gamma \Rightarrow \varphi(t), \Delta \;\; \text{(height} < \alpha)}% {\Gamma \Rightarrow \forall x\varphi(x), \Delta} \;\;\;\; (\omega)\text{-rule}$

Infinite proof trees are indexed by ordinal heights below some fixed bound, such as $\omega_1^{CK}$ . Cut elimination proceeds by double induction on the syntactic complexity of the cut formula and the ordinal height of the derivations.

Regularity and Structural Constraints

Unlike many historical systems, recent frameworks allow unrestricted side-contexts in rules for predicates like truth, with cut-elimination recovering all standard properties (including the invertibility of rules and admissibility of contraction) via appropriate initial sequent restrictions (Nicolai, 2020).

3. Cut-Elimination Methodologies and Ordinal Analysis

A recurring theme in infinitary cut elimination is the use of transfinite induction on a complexity measure—often based on the syntactic complexity (cut-rank), the number or depth of fixed-point unfoldings, and/or a proof height valued in the ordinals. For instance:

In systems of transparent truth, one introduces the τ (T-complexity) measure, which counts the number of truth rule applications above any formula occurrence (Nicolai, 2020).
In arithmetic and set-theoretic systems, ordinal assignments (with notation systems up to $\varepsilon_{\rho_0+1}$ ) direct the cut-elimination process, ensuring every rewriting step strictly lowers the assigned ordinal (Arai, 2018, Walsh, 2021).

Table: Summary of Complexity Measures in Infinitary Cut-Elimination

System	Complexity Measure	Ordinal Bounds
Transparent truth	Cut-rank, τ (T-complexity)	hyperexponential ( $2_m^n$ ) or $\omega_1^{CK}$
KP set theory	Formula degree, proof height	$\varepsilon_{\rho_0+1}$
Arithmetic (ω-rules)	Cut-rank, height	Below $\varepsilon_0$
μ-calculi	Thread depth, fixed-point unfoldings	ω, controlled by trace condition

Progress in cut elimination is thus well-directed, ensuring the process is terminating (well-founded), even when transforming infinite, non-wellfounded derivations.

4. Non-wellfounded, Cyclic, and Ill-founded Proof Systems

Infinitary cut-elimination has been extended to "ill-founded" or "non-wellfounded" systems, including cyclic proofs and systems where infinite proof trees are valid only if they satisfy a global progressivity (trace) criterion (Curzi et al., 1 Feb 2026, Hori et al., 2023, Cerda et al., 9 Oct 2025). These criteria typically ensure that, along every infinite branch of a proof, an infinitary operation (e.g., greatest fixed point unfolding, progress along a thread) occurs infinitely often.

A central methodology is to adapt Tait–Girard reducibility candidates or to use topological properties (internal closure in the space of branches of the proof tree) to guarantee that cut elimination preserves the correctness criterion (Curzi et al., 1 Feb 2026).

In many such systems, cut elimination is realized as a sequence of reductions (possibly transfinite), with the essential property that every transfinite reduction sequence can be compressed to a countable one (the compression property) (Cerda et al., 9 Oct 2025). This justifies the equivalence of coinductive cut-elimination and classic infinitary rewriting.

5. Semantic and Model-Theoretic Consequences

Cut elimination in infinitary settings underpins fixed-point semantics, completeness, and correspondence results. For transparent truth systems, cut-free infinitary provability precisely characterizes the minimal Kripke fixed point of the semantic operator defining the truth predicate (Nicolai, 2020). For cyclic and non-wellfounded proofs, equivalence with cut-free infinitary proofs yields completeness w.r.t. intended models of fixed-point logics and convergence of Herbrand-style expansions (Hori et al., 2023, Cerna et al., 2016).

In particular, the existence of a cut-free infinitary proof with ordinal-bounded height corresponds to the inclusion of the coded sentence in the fixed point under the relevant semantic operator.

6. Complexity, Speedup, and Limitative Phenomena

Systems admitting infinitary cut elimination often exhibit significant proof-theoretic speedup relative to their finitary or purely syntactic counterparts. For instance, schematic CERES approaches may collapse infinitely many distinct cut instances into a single cut-free schema of polynomial size, yielding non-elementary speedup (Cerna et al., 2016). Complexity bounds on the length or height of cut-free proofs are typically given in terms of hyperexponential (e.g., $2^n$ ), primitive-recursive, or higher-ordinal (e.g., $\omega_1^{CK}$ ) functions, depending on the logical fragment and presence of infinitary rules (Nicolai, 2020, Walsh, 2021, Arai, 2018, Pischke, 2017).

A notable limitative result is that, despite cut-elimination, the derivability problem for certain infinitary systems (e.g., those allowing non-local contraction in subexponentials) is $\Pi^1_1$ -complete (i.e., analytic but non-hyperarithmetical) and the closure ordinals of the relevant operators are maximal among admissible induction principles (Kuznetsov et al., 2020).

Recent research generalizes infinitary cut elimination across a variety of proof-theoretic settings:

Schematic cut elimination combines infinitary analyses with schematic representations of proof families, enabling algorithmic extraction of Herbrand systems and uniform elimination of families of high-complexity cuts (Cerna et al., 2016).
Automata-theoretic approaches leverage bounded stack-height or automaton-recognizable trace conditions to recover decidability and algorithmic cut elimination in otherwise undecidable infinitary systems (Baelde et al., 2020).
Cut elimination in modal and fixed-point logics: Custom cyclic calculi for modal μ-calculus, systems with master modalities, and systems for infinitary action logics all instantiate the core transfinite, progress-based cut elimination paradigm (Afshari et al., 13 Oct 2025, Shamkanov, 2023, Miranda et al., 5 May 2025, Kuznetsov et al., 2020).

A consistent theme in these developments is the unification of proof-theoretic, semantic, and computational phenomena through sophisticated techniques for eliminating cut in the presence of infinitary and coinductive constructs.

References

(Nicolai, 2020) Cut elimination for systems of transparent truth with restricted initial sequents
(Cerna et al., 2016) Schematic Cut elimination and the Ordered Pigeonhole Principle
(Baelde et al., 2020) Bouncing threads for infinitary and circular proofs
(Arai, 2018) Cut-elimination for $\omega_{1}$
(Curzi et al., 1 Feb 2026) Making progress: Reducibility Candidates and Cut Elimination in the Ill-founded Realm
(Hori et al., 2023) Cut elimination for propositional cyclic proof systems with fixed-point operators
(Cerda et al., 9 Oct 2025) Compression for Coinductive Infinitary Rewriting: A Generic Approach, with Applications to Cut-Elimination for Non-Wellfounded Proofs
(Pischke, 2017) On Infinitary Gödel logics
(Kuznetsov et al., 2020) Infinitary Action Logic with Exponentiation
(Afshari et al., 13 Oct 2025) Cut-elimination for the alternation-free modal mu-calculus
(Shamkanov, 2023) On structural proof theory of the modal logic K+ extended with infinitary derivations
(Miranda et al., 5 May 2025) Cut elimination for a non-wellfounded system for the master modality
(Acclavio et al., 2023) Infinitary cut-elimination via finite approximations (extended version)
(Walsh, 2021) Reflection ranks via infinitary derivations