Skip to content

Non-standard model of arithmetic

A non-standard model of arithmetic satisfies first-order Peano arithmetic while containing elements beyond every standard numeral.

Version
v2 · 2026-10-03 · History
Domain-specific #
13469
Domain group
Formal Sciences
Origin domain
Mathematics
Subdomain
Models of Arithmetic → Mathematics
Aliases
Nonstandard Model of Arithmetic

Core Idea

A non-standard model of arithmetic satisfies the first-order axioms of Peano arithmetic (PA) yet is not isomorphic to the intended natural-number structure N. It contains the familiar initial segment denoted by the finite numerals 0, S0, SS0, and so on, but also an element larger than every one of those numerals. Such an element is “infinite” only from our external comparison with the standard segment; inside the model it is an ordinary element obeying that model's arithmetic.[1][2]

The existence argument exposes a precise limit of first-order description. PA's induction schema covers each formula in its language, not every externally describable subset of a model. A new constant c can be required to exceed every standard numeral, one sentence per numeral. Each finite subset of these demands is satisfiable in N by choosing a large enough natural number; first-order compactness then gives a model satisfying them all. No standard natural can be that model's c.[1][3]

Structural Signature

Sig role-phrases:

  • First-order arithmetic theory: axioms for zero, successor, operations and formula-by-formula induction.
  • Standard numeral segment: interpretations of each externally finite successor term, forming an initial segment.
  • Nonstandard witness: a model element outside that segment, above every named numeral.
  • Model-construction principle: compactness or an ultrapower makes the enlarged structure coherent.
  • External limitation: order-type, computability and second-order semantics are facts about the model from outside, not extra ordinary natural numbers inside it.[1][2][4]

The distinction between “every number in the model” and “every standard numeral we can write externally” is decisive. PA proves many universal facts inside every model. It does not provide a first-order formula whose extension is exactly the externally standard elements in every nonstandard model. Treating the external list 0,1,2,… as one internal predicate would conceal the compactness construction.[4][2]

What It Is Not

The nonstandard model is not N with a very large ordinary finite number named c. A finite batch of inequalities permits that; the full infinite family does not. Nor is it a claim that PA's arithmetic equations fail. Each model satisfies them, with its own interpretations of addition and multiplication. The novelty is model non-isomorphism, not an ordinary arithmetic counterexample.[1]

It is also not a model of full second-order PA under standard semantics. There, induction quantifies over all subsets of the universe; MIT's logic notes state the resulting theory is categorical, meaning every full model is isomorphic to N. “Second-order” under Henkin semantics has a restricted range of sets and does not inherit that categoricity; the semantic qualifier is essential.[4][5]

Scope of Application

These models are studied in mathematical logic and foundations to test what first-order axioms determine. The compactness construction gives existence without enumerating an actual addition and multiplication table for a chosen countable model. A different construction, an ultrapower N^N/U by a nonprincipal ultrafilter, directly displays a witness: the class of the identity sequence (0,1,2,…) exceeds every constant-sequence numeral. Łoś's theorem makes the ultrapower satisfy every first-order sentence true in N.[1]

The order structure has a striking regularity. Kach's original mathematical exposition states Henkin's theorem: after the initial standard N segment, each nonstandard element sits in a successor/predecessor block ordered like Z, and the blocks themselves form a dense order without endpoints. For a countable nonstandard model, that quotient order is Q, so its order type is N + Z·Q. This describes the Order, not the full addition/multiplication structure; different countable models can share that order type.[2]

Clarity

To see compactness without mistaking a finite bound for a nonstandard number, take the demands c > 0, c > 1, …, c > 100. The standard interpretation c = 101 satisfies that finite batch. For any finite batch, choose a larger standard value. Compactness converts this finitely satisfiable set into a model of the entire infinite theory, in which no numeral n can equal or exceed c. This model may be chosen countable by the downward Löwenheim–Skolem theorem, but compactness alone does not present a computable table of its operations.[1][2]

In the ultrapower, the embedded standard number n is the class of the constant sequence (n,n,n,…). Let e(i)=i. The indices where e(i)>n form a cofinite set, which lies in every nonprincipal ultrafilter on N. Therefore the class [e] exceeds [constant n] for every standard n. Schimmerling also computes that [e]+[constant k]=[i↦i+k] and that [i↦⌊i/2⌋] is another nonstandard element below [e]. This is an executed structure, not just a name for an unspecified “huge integer.”[1]

Manages Complexity

The model perspective separates consistency, cardinality, order and computability. Compactness establishes a model; Löwenheim–Skolem can make one countable; Henkin's order theorem describes its order type; Tennenbaum's theorem rules out a computable presentation of a countable nonstandard PA model. None of those conclusions is interchangeable. In particular, “countable” means there is an external bijection with N, not that the model's operations can be computed from that enumeration.[2][3]

Hermes and Kirst's original research revisits and mechanizes Tennenbaum's theorem. The safe formulation here is that no countable nonstandard model of PA has a computable presentation with both arithmetic operations. This theorem should not be asserted for the ultrapower just described without establishing countability, nor confused with the simple fact that standard N has computable operations.[3]

Abstract Reasoning

The two constructions isolate the same witness by different means. Compactness says that every finite list of lower bounds can be met and so the whole first-order theory has a model. The ultrapower exhibits an element whose sequence eventually exceeds each fixed numeral, with the ultrafilter turning “eventually” into truth in the quotient. Both depend on first-order semantics and both preserve ordinary arithmetic axioms; neither makes a standard finite natural larger than all finite naturals.[1][3]

One can also reason from a nonstandard element a to its Z-block: a−2, a−1, a, a+1, a+2, and so on are all nonstandard for every external finite step. A distinct, much smaller nonstandard block can be represented in the ultrapower by [⌊i/2⌋]. Between blocks, further blocks occur. The resulting order picture helps intuition, but it cannot recover the multiplication table; order-isomorphic models need not be isomorphic as arithmetic structures.[1][2]

Knowledge Transfer

The construction can be used to diagnose limits of first-order axiomatization elsewhere, but the PA-specific order and Tennenbaum results do not automatically transfer to arbitrary theories. The same ultrapower technique has applications in nonstandard analysis, yet this entry is about models of arithmetic; real-number infinitesimals require a different base structure and additional claims.[1]

Second-order contrast needs the same care. Full second-order PA excludes nonstandard models by quantifying over every subset, but that full semantics lacks the first-order compactness/completeness package. Koch and Kirst's original formalization contrasts full-semantics incompleteness with Henkin-semantics completeness and compactness. One cannot simultaneously borrow full categoricity and Henkin compactness as though they were properties of a single semantics.[4][5]

Examples

  1. The compactness model. Form the first-order theory PA ∪ {c > n : n is a standard numeral}. A finite subset has a standard model by interpreting c above its largest mentioned numeral; compactness supplies a model of all constraints, and Löwenheim–Skolem yields a countable one if desired. Mapped back: theory = PA; standard segment = numeral interpretations; witness = c above every numeral; construction = finite satisfiability plus compactness; external limitation = a countable nonstandard choice has N+Z·Q order and cannot have computably presented PA operations. The last two facts are separate theorems, not consequences of the finite check alone.[1][2][3]

  2. Schimmerling's ultrapower. Let U be a nonprincipal ultrafilter and form N^N/U. Standard n embeds as [i↦n]; the identity class [i↦i] is above every one because {i : i>n} is cofinite and hence in U. Mapped back: theory = all first-order truths of N via Łoś; standard segment = constant-sequence classes; witness = identity-sequence class; construction = ultrafilter quotient; external limitation = this construction is not claimed to be countable or computably presented. Schimmerling's [⌊i/2⌋] example shows a distinct lower nonstandard block, not merely a finite predecessor of [i↦i].[1]

Structural Tensions

First-order compactness versus categorical capture of the intended naturals. First-order PA has a formula-by-formula induction schema and the compactness theorem, enabling systematic model construction—but admits nonstandard models. Full second-order induction ranges over every subset and categorically selects N, but loses the first-order compactness/completeness package. Restricting second-order quantifiers to Henkin ranges recovers first-order-style metalogical tools while forfeiting the full categoricity claim. Diagnostic: is the argument using first-order, full second-order or Henkin semantics, and which of compactness or categoricity does it require?[4][5]

Structural–Framed Character

The relation between numeral segment, nonstandard witness and model satisfaction is a mathematical structure independent of human preference. Its evaluative importance arises in foundations, where one asks how completely axioms identify the intended N. The object does not depend on social practice for its existence, although formal languages and proof systems are human-developed tools. The vocabulary travels to ultrapowers because first-order satisfaction is preserved, not because “infinite” sounds similar. Importing Tennenbaum's countability limit into an uncountable ultrapower, or full second-order categoricity into Henkin semantics, would be an invalid transfer. Its character: a formally structural model-theoretic object with exact semantic and cardinality boundaries.[1][3][4]

Structural Core vs. Domain Accent

The skeleton is a carrier with interpreted arithmetic operations and relations satisfying first-order PA while containing a witness beyond every standard numeral. The live Mathematical Structure entry is the genus of carrier, typed operations and axioms; the syntactic theory is not itself this model. The domain mechanism is first-order arithmetic, compactness/ultrapower construction and internal addition and multiplication; the two constructions are accents. The named entry fails the prime bar because its counterintuitive properties—Z-blocks, Tennenbaum's limit and second-order contrast—depend on PA and its semantics, not a generic “extra element” pattern. A prime about axiomatic underdetermination would need independently sourced unlike-domain examples and cannot inherit these arithmetic theorems.[2][3]

This entry is a kind of Mathematical structure.

The live Mathematical Structure entry is the strict parent: each non-standard first-order PA model has a carrier and interpreted operations/relations satisfying axioms, whereas many mathematical structures are not PA models. Compactness is an explanatory metatheorem here, not a separate parent.

Relationships to Other Abstractions

Local relationship map for Non-standard model of arithmeticParents appear above the current abstraction, mutual partners to the right, and children below. Node labels state whether each abstraction is prime or domain-specific; colors identify relation types.Non-standard modelof arithmeticDOMAINDomain-specific abstraction: Mathematical structure — is a kind ofMathematicalstructureDOMAIN

Current abstraction Non-standard model of arithmetic Domain-specific

Parents (1) — more general patterns this builds on

  • Non-standard model of arithmetic is a kind of Mathematical structure Domain-specific

    A non-standard PA model is a mathematical structure.

Hierarchy path (1) — routes to 1 parentless root

Neighborhood in Abstraction Space

Non-standard model of arithmetic sits in a moderately populated region (58th percentile for distinctiveness): it has near-neighbors but no dense thicket of look-alikes.

Family — Formal Models & Logical Foundations (33 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-08

Not to Be Confused With

  • The standard N with a merely large finite number interpreting c for finitely many conditions.
  • A violation of PA's ordinary equations; nonstandard models satisfy the theory.
  • A full second-order PA model under standard semantics, which is categorical.
  • A countable set with necessarily computable arithmetic operations, or an ultrapower assumed countable without proof.[3][4]

References

[1] Ernest Schimmerling, Basic and Intermediate Logic, ch. 5 “Arithmetic,” §§5.1–5.3, original author-hosted mathematical exposition, including compactness and the worked N^N/U construction. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k ↩l ↩m

[2] Asher M. Kach, Non-Standard Models of Arithmetic (2004), original author slides, pp. 2–4, existence and Henkin order-type theorem. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i

[3] Marc Hermes and Dominik Kirst, “An Analysis of Tennenbaum's Theorem in Constructive Type Theory” (2024), original research and mechanized proof, Introduction and theorem. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h

[4] MIT OpenCourseWare, Peano Arithmetic, Logic II lecture notes (2004), pp. 1, 13–15, first-order induction schema and full second-order categoricity. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g

[5] Mark Koch and Dominik Kirst, “Undecidability, Incompleteness, and Completeness of Second-Order Logic in Coq”, original research/formalization, abstract distinguishing full and Henkin semantics. registry ↩a ↩b ↩c