Proof systems and computability can look like separate worlds. Proof theory studies derivations, cut elimination, normalization, and the shape of formal reasoning. Computability studies what can be effectively decided, computed, or reduced to something else. The bridge between them is one of the deepest unifying ideas in logic and foundations: proof is a form of computation, and computation can be organized as a proof search problem.
This bridge is not only philosophical. It is technical and concrete. It lets you translate questions about provability into questions about algorithms and vice versa. It also clarifies why certain proof systems are powerful, why certain logics admit program extraction, and why complexity shows up inside proof theory.
Premium Gaming TV65-Inch OLED Gaming PickLG 65-Inch Class OLED evo AI 4K C5 Series Smart TV (OLED65C5PUA, 2025)
LG 65-Inch Class OLED evo AI 4K C5 Series Smart TV (OLED65C5PUA, 2025)
A premium gaming-and-entertainment TV option for console pages, living-room gaming roundups, and OLED recommendation articles.
- 65-inch 4K OLED display
- Up to 144Hz refresh support
- Dolby Vision and Dolby Atmos
- Four HDMI 2.1 inputs
- G-Sync, FreeSync, and VRR support
Why it stands out
- Great gaming feature set
- Strong OLED picture quality
- Works well in premium console or PC-over-TV setups
Things to know
- Premium purchase
- Large-screen price moves often
Proofs as objects and procedures
A proof system gives rules for deriving judgments. A computability model gives rules for transforming input into output. The bridge begins when you make proofs into explicit objects:
- A proof is a finite tree of rule applications
- A proof can be encoded as a finite string
- A proof can be checked mechanically by verifying each rule instance
Once you see proofs as finite objects, two computational notions appear immediately:
- Proof verification is an effective procedure
- Proof search is a computational task whose difficulty depends on the system
This is the first link: a proof is a certificate, and a proof system is a language of certificates.
The Curry Howard correspondence: propositions as types
The most famous bridge is the Curry Howard correspondence. In its core form:
- Propositions correspond to types
- Proofs correspond to terms
- Normalization corresponds to computation
In a constructive setting, a proof of an implication `P → Q` is a function that takes a proof of `P` and returns a proof of `Q`. This is not metaphor. In a typed \lambda calculus, an inhabitant of a function type is literally a function term.
The consequence is that a constructive proof carries computational content:
- From a proof of an existential statement, you can often extract a witness
- From a proof that a function exists, you can often extract an algorithm
This explains why type theory has become a foundation for verified software. It is not because it is fashionable. It is because the proof calculus is already a programming calculus.
Sequent calculus and computation by normalization
Another form of the bridge appears in sequent calculi. The sequent calculus makes the structure of proofs explicit through left and right rules and through the cut rule.
Cut elimination is a theorem that any proof using cuts can be transformed into a cut free proof. The cut free proof has the subformula property: every formula in the proof is a subformula of the goal and hypotheses.
Computationally, cut elimination acts like evaluation:
- A cut is like applying a lemma or a function call
- Eliminating a cut is like inlining or reducing that call
- The resulting cut free proof is like a normal form
This viewpoint makes proof transformations resemble program transformations. It also makes the search space visible. Cut free proofs are constrained, which is good for decidability results in certain fragments, but they can also be exponentially large, which is why proof complexity becomes relevant.
Proof search and decision procedures
For some logics, proof search is decidable. For others, it is not. The bridge with computability is:
- A decision procedure for validity is a proof search algorithm that always terminates
- Undecidability results can be proved by reducing a known undecidable computational problem to proof search
Classical first order logic has a complete proof system, but validity is undecidable. This means:
- If a statement is valid, there is a proof
- There is no algorithm that will always determine validity for arbitrary statements
That situation is a precise reflection of incompleteness style limits. Completeness provides existence of proofs, undecidability denies uniform search termination guarantees.
In many sublogics, the situation improves. For example:
- Propositional logic has decidable validity, and proof search can be automated
- Certain modal logics have decidable proof search with strong complexity bounds
- Certain fragments of first order logic have decidable satisfiability via syntactic restrictions
The bridge tells you what to look for: the shape of formulas that restricts the proof search tree.
Computability enters proof theory through representability
Many limit theorems in proof theory depend on the fact that arithmetic can represent computation. Once a theory can represent basic computable relations, you can express:
- The statement that a computation halts
- The statement that a proof exists for a given formula
This makes reductions natural:
- If you can express halting, you can encode undecidability into provability questions
- If you can express provability, you can encode incompleteness into arithmetic truth questions
In practice, the representability lemmas are the bridge components: they connect the syntax of computation to the syntax of proof inside a theory.
Proofs as certificates and the geometry of complexity
From a computational viewpoint, a proof is a certificate that can be checked. This brings in complexity almost automatically.
Two complexity measures matter:
- The time to verify a proof of length n
- The minimal length of a proof of a given statement in a given system
The second measure leads to proof complexity. It studies:
- Lower bounds on proof size for families of tautologies
- Separations between proof systems by comparing shortest proofs
This is not merely about speed. It is about the structure of reasoning. Different proof systems allow different compression of arguments.
A practical mental model is:
- Some proof systems are expressive but have large search spaces
- Some proof systems are restrictive but make certificates easier to find
- Extensions like additional rules or axioms can be understood as adding computational power to the certificate language
Program extraction: turning proofs into algorithms
In constructive settings, program extraction is a disciplined method:
- Prove a specification theorem in a constructive logic
- Extract an algorithm from the proof term
- Prove that the extracted algorithm meets the specification
This is a direct manifestation of the propositions as types principle. The extracted program is the proof term interpreted as code.
A key distinction is between:
- Proof relevant content, which becomes executable
- Proof irrelevant content, which is erased in extraction
Modern systems formalize this distinction with techniques like implicit arguments, erasure, or separate universes for propositions and data. The foundational point is that the proof system is also a calculus of programs.
Realizability and models that remember computation
Realizability provides another bridge. It interprets logical formulas by associating them with computational witnesses, often called realizers.
The idea is:
- A formula is true if there is a computational object that realizes it
- Logical connectives correspond to computational constructions
- Existence corresponds to providing a witness with evidence
Realizability models show that constructive logics are consistent by building semantic universes in which truth is witnessed by computation.
This also gives refined insights:
- Some classical principles do not have realizers in certain computability frameworks
- Some principles become realizable with added computational features, such as choice operators
Realizability therefore connects proof principles to computational resources.
A table map: proof artifacts and computational artifacts
| Proof theory notion | Computability analogue | What the translation buys |
|—|—|—|
| derivation tree | computation trace | explicit structure for verification |
| normalization | evaluation | meaning by reduction to normal forms |
| cut elimination | inlining or \beta reduction | subformula property and canonical proofs |
| proof search | algorithmic search | decidability and complexity analysis |
| proof size | certificate size | lower bounds and system comparisons |
| constructive proof | program | witness extraction and verified computation |
This table is not an analogy list. It is a guide for navigating papers. When you see one side, you can predict the other side will appear.
Why the bridge matters for foundations
Foundations is not only about what exists. It is about what can be certified. Proof systems provide certificates. Computability tells you which certificates can be found by algorithms and which cannot.
This explains several recurring themes:
- Completeness theorems guarantee existence of proofs, but do not guarantee search feasibility
- Undecidability results show there is no uniform procedure for finding proofs in full generality
- Constructive systems embed computation inside proof, enabling extraction and verification
- Strength comparisons between systems often correspond to differences in the computational content they can express
The bridge is therefore the grammar of modern foundational work.
Interactive proof assistants and machine checked computation
The bridge becomes visible in practice when you use an interactive prover. A proof assistant turns the rules of a proof system into a kernel that checks derivations. Tactics and automation then become search procedures, guided by human intent. On the computational side, this looks like a certified pipeline:
- You describe a specification in a formal language
- The system constructs a proof term, partly by automation and partly by guidance
- The kernel verifies the proof term, producing a trustworthy certificate
This workflow shows why foundations cares about small trusted kernels, explicit proof objects, and extraction. It is not only about philosophy. It is about building a chain of responsibility from claim to certificate.
Closing perspective
Proof systems and computability are two views of one underlying structure: finite syntactic objects manipulated by effective procedures under constraints. Seeing the bridge makes many results feel inevitable rather than surprising. A proof is a certificate, proof transformations are computations, and computation can often be framed as proof search.
Once you read logic and foundations with that unity in mind, the subject becomes less like a set of disconnected subfields and more like a coherent theory of certification, computation, and the boundaries that constrain both.
Books by Drew Higgins
Christian Living / Encouragement
God’s Promises in the Bible for Difficult Times
A Scripture-based reminder of God’s promises for believers walking through hardship and uncertainty.
Leave a Reply