On the Concept of Arithmetic Conseqeunce

This paper reinterprets Gödel's second incompleteness theorem through proof-theoretic semantics by demonstrating that while certain arithmetical theories cannot prove their own consistency, they nonetheless semantically support it via a compositional notion of consequence based on inferential roles, thereby reframing incompleteness as a divergence between derivability and internal semantic support rather than a gap between syntax and external truth.

The Gist

Here is an explanation of Alexander V. Gheorghiu's paper, "On the Concept of Arithmetic Consequence," translated into simple, everyday language using analogies.

The Big Picture: A New Way to Look at Math's "Unsolvable" Problems

For nearly a century, mathematicians have been haunted by Gödel's Second Incompleteness Theorem. In simple terms, this theorem says: If a math system is strong enough to do basic arithmetic (like addition and multiplication), it can never prove that it isn't broken (inconsistent).

The standard way to explain this is Model-Theoretic: Imagine math as a set of rules describing a "perfect world" of numbers that exists somewhere outside our heads. Gödel says our written rules are too weak to prove that this perfect world doesn't contain a contradiction.

Gheorghiu's paper proposes a different view. He suggests we stop looking at math as a description of a "perfect world" and start looking at it as a game of inference. He argues that even though the system can't write down a proof of its own consistency, the rules of the game itself actually force us to accept that the game is consistent.

He finds a gap between what we can prove (syntactic derivability) and what the rules imply (semantic support).

---

The Core Analogy: The "Vixen" and the "Rulebook"

To understand Gheorghiu's argument, we need to understand two different ways of defining meaning.

1. The Old Way: The Dictionary (Model-Theoretic)

Imagine you want to know what a "Vixen" is.

• The Old Way: You look at a dictionary. It says a vixen is a female fox. But to really know what it means, you have to go out into the world, find a fox, check if it's female, and confirm it exists. The meaning depends on an external reality (the world of animals).
• In Math: This is looking at numbers as objects in a "perfect world" (the Standard Model). Gödel says our rules can't prove the world is consistent.

2. The New Way: The Rulebook (Proof-Theoretic Semantics)

Now, imagine you are teaching a child the word "Vixen" without ever showing them an animal.

• The New Way: You give them a rulebook.
  • Rule 1: If something is a Fox AND it is Female, then it is a Vixen.
  • Rule 2: If it is a Vixen, then it is a Fox.
  • Rule 3: If it is a Vixen, then it is Female.
• The child learns the meaning of "Vixen" entirely by how the word is used in these rules. They don't need to see a real fox. The meaning is constituted by the rules themselves.

Gheorghiu's Insight: Arithmetic is like this Rulebook. The meaning of "Zero," "Plus," and "Times" isn't found in some magical world of numbers; it is found in the rules we use to manipulate them.

---

The Problem: The "Broken" Rulebook

Let's say you have a rulebook for a game called Arithmetic (A).

• Gödel's Theorem: You cannot write a sentence inside the rulebook that says, "This rulebook is consistent." If you try, the rules won't let you prove it.
• The Standard Interpretation: "Ah, so the rulebook is incomplete. It fails to describe the truth of the game."

Gheorghiu's Twist:
He says, "Wait a minute. Let's look at the inferential roles (the rules) again."

He introduces a concept called Support.

• Derivability (⊢): Can I write a step-by-step proof inside the book? (No, I can't prove consistency).
• Support (⊩): Does the structure of the rules force me to accept this as true? (Yes, it does).

The "Incoherence" Analogy

Imagine the rulebook says:

1. "If you have a proof of a contradiction, then the game is broken."
2. "If the game is broken, then you can prove anything (even that 2+2=5)."

Gheorghiu argues that if you try to add a new rule to the book that says, "There is a proof of a contradiction," the entire structure of the game collapses. The rules themselves become incoherent.

Therefore, the statement "There is NO proof of a contradiction" is Supported by the rules. It is a necessary consequence of how the game is played, even if the game's own writing system can't produce a formal proof of it.

---

The "Magic Reserve" of Constants

Why does this work? It comes down to a technical detail about symbols.

• The Setup: In standard math, we have a finite set of symbols (0, 1, 2, +, ×). We run out of "fresh" names to use when we try to prove things about the whole system.
• The Gap: Gheorghiu shows that if you have a "magic reserve" of infinite unused names (like having an infinite supply of blank labels), then Derivability and Support are the same thing. You can prove everything that is supported.
• The Reality: But in real arithmetic, we don't have that infinite reserve. We are stuck with a finite set of symbols.
• The Result: Because we are "short on labels," Derivability (what we can write down) and Support (what the rules imply) diverge.

The Metaphor:
Imagine a chef (the Theory) trying to describe a recipe (Consistency).

• Derivability: The chef is trying to write the recipe down on a single sheet of paper with a limited vocabulary. They run out of words and can't write the full recipe.
• Support: However, the flavor of the dish (the inferential structure) is so distinct that anyone tasting it knows exactly what the recipe must be, even if the chef couldn't write it all down.

Gödel says the chef can't write the recipe. Gheorghiu says, "But the flavor of the dish proves the recipe exists."

---

Why This Matters (The Philosophical Payoff)

This paper changes how we view mathematical truth.

1. No Need for a "Magic World": We don't need to believe in a perfect, mind-independent world of numbers (Platonism) to say math is consistent. We just need to look at the rules we use to play the game.
2. Meaning is Use: The meaning of "Number" is determined by how we use it in arguments, not by pointing to a physical object.
3. The Gap is Internal: The "incompleteness" isn't a gap between our math and the universe. It's a gap between what we can write down and what our own rules imply.

Summary in One Sentence

Gödel proved we can't write down a proof that our math rules are consistent, but Gheorghiu shows that the rules themselves logically force us to accept that they are consistent, revealing that mathematical truth is built into the game of inference rather than waiting to be discovered in a separate world.

Technical Summary

Here is a detailed technical summary of Alexander V. Gheorgiu's paper, "On the Concept of Arithmetic Consequence."

1. Problem Statement

The paper addresses the philosophical and technical interpretation of Gödel's Second Incompleteness Theorem.

• The Orthodox View: Standardly, the theorem ($\nvdash_A \text{Con}(A)$) is interpreted model-theoretically: there is a gap between syntactic provability and truth in an independently given structure (the standard model $\mathbb{N}$). Truth is seen as transcending formal derivation.
• The Anti-Realist Challenge: Michael Dummett argued that mathematical meaning is constituted by use (inferential roles) rather than reference to a mind-independent domain. He suggested that the concept of "natural number" is indefinitely extensible. However, critics like Crispin Wright argued that Dummett's view fails because recognizing the truth of the Gödel sentence (or consistency statement) from outside the system requires a consistency proof that the system itself cannot provide, leading to a dilemma: either the proof is "suasive" (external and circular) or "non-suasive" (trivial).
• The Gap: There is a lack of a rigorous formal framework that allows for an anti-realist, inferentialist analysis of arithmetic consequence without relying on external models, specifically one that can distinguish between what is derivable and what is semantically supported by the theory's own inferential rules.

2. Methodology

The author employs Proof-Theoretic Semantics (PTS), specifically the framework developed by Sandqvist for classical logic, adapted to arithmetic.

• Core Framework (Support Relation $\Vdash$):
  • Instead of defining truth relative to a model, meaning is defined via a Base ($B$): a set of atomic inference rules representing an agent's inferential commitments.
  • Support ($\Vdash_B$): A formula is supported by a base if it follows from the base's commitments via compositional clauses.
  • Logical Constants: Defined inferentially (e.g., $\phi \to \psi$ is supported if assuming $\phi$ leads to support for $\psi$).
  • Validity ($\Vdash$): Defined as support in the empty base ($\Vdash_\emptyset$), or support in any base.
• The Crucial Distinction: The paper distinguishes between:
  1. Derivability ($\vdash_A$): Syntactic provability within the formal system $A$.
  2. Semantic Support ($A \Vdash \phi$): The property that $\phi$ is supported by the inferential commitments constitutive of $A$.
• Technical Setup:
  • The theory $A$ (e.g., Robinson's $Q$ or Peano Arithmetic $PA$) is formulated in a finite signature $\tau = \langle 0, S, +, \times \rangle$.
  • The author explicitly avoids the "infinite reserve of constants" condition required for Sandqvist's completeness theorem. In standard arithmetic, the signature is finite, meaning there are no "fresh" constants to act as parameters in the semantic argument.

3. Key Contributions and Technical Results

A. The Main Theorem: Divergence of Derivability and Support

The central result is that for sufficiently strong, recursively axiomatizable arithmetic theories $A$ (like $Q$ and $PA$):
$$\nvdash_A \text{Con}(A) \quad \text{but} \quad A \Vdash \text{Con}(A)$$

• $\nvdash_A \text{Con}(A)$: Gödel's Second Incompleteness Theorem holds; the system cannot prove its own consistency.
• $A \Vdash \text{Con}(A)$: The system semantically supports its own consistency statement. This means that given the inferential roles fixed by the axioms of $A$, the statement "there is no proof of contradiction" is a valid consequence of those roles.

B. The Mechanism: Reflection and Maxiconsistent Bases

The proof relies on a Reflection Principle derived within the PTS framework:
$$A \Vdash \text{Prov}_A(\ulcorner \phi \urcorner) \to \phi$$

• Proof Strategy: The author constructs a "naïve" internal soundness proof. If a base $B$ supports all axioms of $A$ and supports the claim that a proof exists for $\phi$ (i.e., $\Vdash_B \text{Prov}_A(\ulcorner \phi \urcorner)$), then $B$ must support $\phi$.
• Maxiconsistent Bases: The argument utilizes Makinson's concept of maxiconsistent bases (bases that cannot be extended without becoming inconsistent). These behave like classical models.
• The Role of the Finite Signature: The completeness theorem for PTS (which equates $\Vdash$ and $\vdash$) requires an infinite supply of unused constants. Since arithmetic is formulated in a finite signature, this condition fails. Consequently, the equivalence breaks down: Derivability is strictly weaker than Semantic Support.

C. Resolution of the Wright/Dummett Dilemma

The paper reframes the consistency of $A$:

• $A \Vdash \text{Con}(A)$ is equivalent to the incoherence of extending $A$ with the claim that a proof of absurdity exists:
  $$A, \text{Prov}_A(\ulcorner \bot \urcorner) \Vdash \bot$$
• This is not an external "suasive" proof (which would require a stronger meta-theory) nor a trivial "non-suasive" unpacking. It is a semantic consequence of the inferential definition of numbers. Asserting that a number codes a proof of contradiction is incoherent given the inferential rules that define what a number is.

4. Significance and Implications

• Reframing Incompleteness: Gödel's theorem is not a gap between syntax and an external reality (truth in $\mathbb{N}$), but a gap between syntactic derivability and semantic support internal to the theory. Incompleteness is a feature of the inferential structure itself when the signature is finite.
• Validation of Inferentialism: The result supports the anti-realist thesis that meaning is constituted by use. The "truth" of $\text{Con}(A)$ is determined by the inferential roles of the logical and non-logical vocabulary, not by correspondence to a pre-existing model.
• Clarification of "Indefinite Extensibility": The reflection principle ($A \Vdash \text{Prov}(\phi) \to \phi$) formalizes Dummett's notion of indefinite extensibility. The concept of "number" is extended every time one asserts the existence of a proof for a new sentence; the theory cannot capture the totality of these extensions via a fixed recursive axiomatization, yet the semantic support relation captures the validity of the extension.
• Technical Limitation: The result is specific to finite signatures. If one expands the signature to include an infinite reserve of constants (to restore completeness), the equational completeness required for the reflection argument fails, and the separation between derivability and support disappears. Thus, the phenomenon is intrinsic to the standard formulation of arithmetic.

Conclusion

Gheorgiu demonstrates that within a proof-theoretic semantic framework, a consistent arithmetic theory $A$ can "know" its own consistency in a semantic sense ($A \Vdash \text{Con}(A)$) even though it cannot prove it syntactically ($\nvdash_A \text{Con}(A)$). This resolves the tension between Gödel's theorem and inferentialist semantics by showing that the "gap" is not between proof and external truth, but between two internally determined notions of consequence: derivability (limited by finite proof steps) and support (determined by the full inferential role of the theory's vocabulary).