Tarskian truth theories over set theory
Abstract: This work uses mostly model-theoretic methods to establish new proof-theoretic theorems about several axiomatic theories of truth over KP (Kripke-Platek set theory) and stronger theories, especially ZF (Zermelo-Fraenkel set theory).
- Nonclassical truth with classical strength. A proof-theoretic analysis of compositional truth over HYPE (2020)
- Axiomatic theories of supervaluational truth: completing the picture (2024)
- Kreisel-Lévy-type theorems for Kripke-Platek and other set theories (2022)
- Constructing the Constructible Universe Constructively (2022)
- Admissible extensions of subtheories of second order arithmetic (2022)
- Provability Logic: models within models in Peano Arithmetic (2021)
- Reflection algebras and conservation results for theories of iterated truth (2019)
- Power Kripke-Platek set theory and the axiom of choice (2018)
- Classifying the provably total set functions of KP and KP(P) (2016)
- Ordinal analysis of intuitionistic power and exponentiation Kripke Platek set theory (2016)
Summary
- The paper’s main contribution is its systematic framework for extending ZF with a compositional Tarskian truth predicate and establishing deductive closure equivalences.
- It precisely characterizes set-theoretical consequences by linking truth predicates to hierarchies of consistency assertions and internal reflection principles.
- The study demonstrates model-existence and conservativity results through advanced model-theoretic techniques, clarifying the limits and strengths of these truth-theoretic extensions.
Authoritative Essay: "Tarskian Truth Theories over Set Theory" (2604.03825)
Introduction and Scope
The paper provides a systematic, highly technical treatment of Tarskian truth theories formulated over set-theoretic backgrounds, centering on extensions of the Zermelo-Fraenkel system, ZF, with a compositional truth predicate, denoted T. The minimal theory, CT−[ZF], augments ZF with finitely many Tarski-style axioms capturing compositional truth (closure of T under logical connectives and quantifiers). A focal point is CT∗​[ZF], defined as CT−[ZF] further extended with the global reflection principle: "All theorems of ZF are true."
The paper investigates the deductive strength, interpretability, and conservativity of these truth-theoretic extensions, contextualizing them within both existing set-theoretic literature and arithmetical analogues. The results are positioned against classical and contemporary work, such as Montague's Reflection Theorem, Levy's Partial Definability, and Krajewski's conservativity results. Strong emphasis is placed on the model-theoretic and proof-theoretic interplay of truth predicates with foundational schemes of set theory (separation, collection, replacement, induction), and on their analogues in theories of classes (e.g., Gödel-Bernays).
Principal Results
Categorization of Deductive Closures
Theorem A establishes deductive closure equivalence among several truth-theoretic extensions:
- CT∗​[ZF]
- CT−[ZF] plus internal replacement and disjunctive correctness outside
- CT−[ZF] plus internal reflection
Internal versions of replacement, separation, and reflection assert the truth of all respective instances in the sense specified by T. Disjunctive Correctness (DCCT−[ZF]0) ensures that the truth of a finite disjunction entails the truth of one of its disjuncts. The equivalence shows that the addition of global reflection for ZF-theorems is, in effect, matched by the closure of the truth predicate under internal replacement and reflection principles.
Characterization of Set-Theoretic Consequences
Theorem B provides a precise characterization of CT−[ZF]1's CT−[ZF]2-consequences: for any sentence CT−[ZF]3 in the language of set theory,
- CT−[ZF]4 if and only if
- CT−[ZF]5 plus a hierarchy of consistency assertions for logical depth CT−[ZF]6 (`CT−[ZF]7' for all CT−[ZF]8) proves CT−[ZF]9.
This result aligns T0 deductively with the union of all finite consistency extensions of ZFC, paralleling results from arithmetic (e.g., for T1 and T2).
Hierarchical Model-Existence Theorems
Theorem C articulates strict hierarchies among truth-theoretic extensions:
- The existence of a well-founded model of ZF is provable in T3-Separation, but not in T4 alone.
- The existence of a model in the form T5 becomes provable only after further extension with T6-collection.
These results delineate the precise impact of additional set-theoretic axioms extended to formulae mentioning T7, and provide explicit counterexamples separating the strength of systems with and without these extensions.
Conservativity
Theorem D demonstrates that T8 is conservative over T9: every CT∗​[ZF]0-sentence provable in this theory is already provable in ZF. The result is obtained via model-theoretic techniques leveraging elementary end-extensions and satisfaction class construction, aligning with advanced results in arithmetic truth theory.
Technical Innovations and Contrasts
The development and comparison of various truth theories over set backgrounds (CT∗​[ZF]1, CT∗​[ZF]2, CT∗​[ZF]3, etc.) are underpinned by detailed analyses of the internal and external closure properties of the truth predicate. The identification of multiple deductively equivalent axiomatizations (the "many faces" theorem) for CT∗​[ZF]4 enables robust transfer of techniques between model-theoretic and proof-theoretic contexts.
The paper generalizes several classical arithmetical results (e.g., conservativity theorems, reflection principles, speed-up results) to the set-theoretic field, including adaptation of methods for constructing full satisfaction classes, and the use of cuts and hierarchies (depth, CT∗​[ZF]5) reminiscent of approaches in arithmetic. Of particular note is the thorough mapping between truth theories over ZF and their natural analogues in the Gödel-Bernays class theory context, establishing mutual interpretability and aligning their set-theoretic consequences.
Explicit counterexamples and constructions elucidate fine distinctions (e.g., pointwise definable minimal models, Paris models, models with or without access to well-foundedness), and demonstrate the boundaries of provability among the hierarchy of truth-theoretic extensions.
Strong Numerical Results and Bold Claims
- Deductive closure equivalence: Multiple truth-theoretic systems (CT∗​[ZF]6, CT∗​[ZF]7-Repl, etc.) yield identical sets of theorems in their extension.
- Set-theoretical consequences: CT∗​[ZF]8 has precisely the CT∗​[ZF]9-consequences as ZFC extended by all consistency assertions of finite logical depth.
- Unprovability of CT−[ZF]0-model existence in CT−[ZF]1: Despite proving the consistency of ZF, CT−[ZF]2 does not ascertain the existence of an CT−[ZF]3-model of ZF, providing a strict separation in strength.
- Conservativity of collection-extended truth theory: Adding collection to compositional truth over ZF, even with the truth predicate, does not increase deductive power in the set language.
Implications and Speculation
Theoretical Implications
The results obtained establish a robust taxonomy of truth-theoretic extensions of ZF, clarifying the impact of compositional truth on foundational schemes and set-theoretic strength. The equivalence of deductive closure among several systems provides a canonical framework for understanding the interaction of reflection and replacement with truth. The characterization of CT−[ZF]4-consequences in terms of consistency hierarchies precisely situates truth-theoretical strength relative to traditional set-theoretic extensions.
Moreover, the paper demonstrates that the addition of global reflection (in the form "all theorems of ZF are true") does not increase the set-theoretical consequences beyond those obtainable by iterative consistency extensions, marking a clear boundary between expressive strength and foundational commitment.
The separation results concerning CT−[ZF]5-models and well-foundedness (provable only with further scheme extension) identify critical loci of strength, informing both set-theoretic and truth-theoretic axiom selection.
Practical Implications and Future Directions
Practically, these findings inform the use of truth predicates in set-theoretic foundations, especially in contexts requiring compositional truth for formalization or philosophical analysis. The conservativity results assure that extending ZF with compositional truth (and even with collection) does not risk unintended consequences in terms of set-theoretic provability. They further clarify the boundaries of such extensions when considering their utility in mathematical logic or foundations.
The identification of mutual interpretability between truth theory and class theory inspires approaches in which axiomatic theories of truth are deployed within or alongside class-theoretic frameworks, facilitating transfer of results and techniques.
Future developments may include further refinement of the speed-up results and interpretability hierarchy, adaptations of the model-theoretic methods to other fragments of ZF or alternate set-theoretic backgrounds, and the exploration of nonstandard models, satisfaction classes, and their impact on truth-theoretic strength.
There also remain several open questions concerning proof-theoretic characterization, the existence of well-founded models relative to further truth-theoretic extensions, and the potential for superpolynomial speed-up in collection-augmented compositional truth theories.
Conclusion
The paper delivers a comprehensive and highly technical analysis of Tarskian truth theories over set theory, rigorously delineating their deductive closures, structural relationships, and implications for set-theoretic foundations. Through model-theoretic and proof-theoretic methods, the work maps the landscape of compositional truth in set contexts, clarifying its conservativity, strength, and limitations. This synthesis of truth-theory and set-theory provides authoritative guidance for further foundational research and informs the deployment of truth predicates in mathematical logic.
Paper to Video (Beta)
No one has generated a video about this paper yet.
Whiteboard
No one has generated a whiteboard explanation for this paper yet.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Open Problems
We haven't generated a list of open problems mentioned in this paper yet.
Continue Learning
- How do the compositional truth predicates influence the deductive strength of ZF extensions?
- What is the role of internal replacement and disjunctive correctness in establishing truth-theoretic equivalences?
- How does the global reflection principle affect the set-theoretic consequences in the paper?
- In what ways can the model-theoretic techniques used here be applied to other foundational systems?
- Find recent papers about Tarskian truth predicates in set theory.
Collections
Sign up for free to add this paper to one or more collections.
Tweets
Sign up for free to view the 1 tweet with 0 likes about this paper.