Natural numbers are the numbers used for counting, ordering, and finite repetition. They are the first numerical structure most readers meet, but their role in mathematics is not merely elementary: they provide the indexing set for [sequences](/page/Sequence), the finite sizes measured by cardinality, the exponents in powers, the iteration counts in algorithms, and the basic objects of number theory. Once the natural numbers are isolated as a structure, familiar procedures such as induction and recursive definition become theorems about that structure rather than informal habits of calculation. Androma follows the convention that the natural numbers start at $1$: \begin{align*} \mathbb{N} = \{1,2,3,\ldots\}. \end{align*} The number $0$ is not a natural number under this convention, although many nearby constructions use $\mathbb{N}_0 := \mathbb{N} \cup \{0\}$ when a zero-based indexing set is needed. ## Definition A counting system needs more than a list of numerals. It needs a first object, a rule for passing to the next object, and axioms that prevent cycles or unexpected extra elements. The Peano viewpoint packages these requirements into a structure from which arithmetic, order, and induction can be recovered. [definition: Natural Number] The set of natural numbers is the set $\mathbb{N}$ whose elements are generated from a distinguished element $1$ by repeated application of a successor operation $S: \mathbb{N} \to \mathbb{N}$, subject to the Peano axioms: $1 \in \mathbb{N}$; $S(n) \in \mathbb{N}$ for every $n \in \mathbb{N}$; $S(n) \ne 1$ for every $n \in \mathbb{N}$; $S(m)=S(n) \implies m=n$ for all $m,n \in \mathbb{N}$; and if $A \subset \mathbb{N}$, $1 \in A$, and $n \in A \implies S(n) \in A$, then $A=\mathbb{N}$. [/definition] The last axiom is the structural source of mathematical induction. It says that no proper subset can contain $1$ and be closed under moving to the next number. Without this condition, a system could contain a genuine copy of the counting numbers together with unrelated extra elements, and induction would not control the whole structure. To define arithmetic from counting rather than assume it in advance, we need a primitive operation expressing only the next step. This operation is the local motion inside the natural numbers: it records how counting advances before addition, multiplication, or order have been built. [definition: Successor Operation] A successor operation on $\mathbb{N}$ is the function $S: \mathbb{N} \to \mathbb{N}$ satisfying the Peano successor axioms $S(n) \ne 1$ for all $n \in \mathbb{N}$ and $S(m)=S(n) \implies m=n$ for all $m,n \in \mathbb{N}$. [/definition] In ordinary decimal notation, $S(n)$ is written $n+1$. That notation already presupposes addition, so the successor map is conceptually prior to addition in an axiomatic construction. Mathematicians often want to compare different constructions of the counting numbers. A model built from finite von Neumann ordinals, a model built from strokes on a page, and a model built abstractly from axioms should count in the same way. The following definition isolates the shared structure so that the underlying set is not mistaken for the mathematical essence. [definition: Peano System] A Peano system is a triple $(N,e,S)$ consisting of a set $N$, an element $e \in N$, and a function $S: N \to N$ such that $S(n) \ne e$ for every $n \in N$; $S(m)=S(n) \implies m=n$ for all $m,n \in N$; and if $A \subset N$, $e \in A$, and $n \in A \implies S(n) \in A$, then $A=N$. [/definition] The most direct model is the counting model itself. It is worth spelling out here because it shows how the abstract successor axioms correspond to everyday numerals before any later arithmetic laws are invoked. [example: Counting Model] Take $N=\{1,2,3,\ldots\}$ and define $S:N\to N$ by $S(n)=n+1$. If $n\in N$, then $n\ge 1$, so $S(n)=n+1\ge 2$; hence $S(n)\in N$ and $S(n)\ne 1$. If $S(m)=S(n)$, then \begin{align*} m+1=n+1. \end{align*} Cancelling $1$ from both sides in the ordinary arithmetic of positive integers gives \begin{align*} m=n. \end{align*} Thus the successor map is injective. Now let $A\subset N$ with $1\in A$ and with the closure property $n\in A\implies S(n)\in A$. Since $1\in A$, closure gives \begin{align*} S(1)=1+1=2\in A. \end{align*} Applying closure again gives \begin{align*} S(2)=2+1=3\in A. \end{align*} Continuing in the same finite way, after $k-1$ successor steps from $1$ we obtain $k\in A$ for each $k\in\{1,2,3,\ldots\}$. Therefore every element of $N$ lies in $A$, so $A=N$. This concrete model satisfies the Peano axioms, so the axioms capture ordinary counting. [/example] The axioms are also designed to reject structures that look partly like counting but contain cycles. Cycles destroy the idea of always moving to a genuinely new next number. [example: A Cyclic Failure] Let $N=\{a,b,c\}$, choose $e=a$, and define $S:N\to N$ by $S(a)=b$, $S(b)=c$, and $S(c)=a$. For a Peano system, the successor operation must satisfy $S(n)\ne e$ for every $n\in N$. Taking $n=c$, the definition gives \begin{align*} S(c)=a. \end{align*} Since $e=a$, this is \begin{align*} S(c)=e. \end{align*} Thus the required condition $S(c)\ne e$ is false, so $(N,e,S)$ is not a Peano system. The failure is exactly cyclic behavior: starting at $e=a$, the successive values are $a$, then $S(a)=b$, then $S(b)=c$, then $S(c)=a=e$. After three successor steps the process returns to the starting element instead of producing a new next element, so this structure cannot model ordinary natural-number counting. [/example] Another failure is subtler. The successor rule may behave correctly on a copy of the natural numbers while leaving extra elements unreachable from the starting point. The induction axiom is exactly what rules this out. [example: Extra Unreachable Element] Let $N=\mathbb{N}\cup\{\omega\}$ with $\omega\notin\mathbb{N}$, let $e=1$, define $S(n)=n+1$ for $n\in\mathbb{N}$, and define $S(\omega)=\omega$. To test the induction axiom, take the subset $A=\mathbb{N}\subset N$. Since $e=1$ and $1\in\mathbb{N}$, we have $e\in A$. Now let $x\in A$. Then $x\in\mathbb{N}$, so the definition of $S$ on elements of $\mathbb{N}$ gives \begin{align*} S(x)=x+1. \end{align*} Because $x+1\in\mathbb{N}$ for every $x\in\mathbb{N}$, this gives \begin{align*} S(x)\in\mathbb{N}=A. \end{align*} Thus $A$ contains $e$ and is closed under $S$. However, $\omega\in N$ by the definition $N=\mathbb{N}\cup\{\omega\}$, while $\omega\notin A$ because $A=\mathbb{N}$ and $\omega\notin\mathbb{N}$. Hence \begin{align*} A\ne N. \end{align*} So there exists a proper subset of $N$ that contains $e$ and is closed under $S$, which is exactly the failure of the induction axiom in the definition of a Peano system. The extra element is unreachable from the starting point: starting at $1$, every successor step stays inside $\mathbb{N}$ because $S(n)=n+1\in\mathbb{N}$ whenever $n\in\mathbb{N}$. Since $\omega\notin\mathbb{N}$, no finite chain of successor steps beginning at $1$ can produce $\omega$. [/example] The next structure needed for counting collections is addition: starting with $n$, count forward $m$ more steps. Defining addition by recursion prevents circularity, because it uses only the already-given successor operation and turns repeated counting into a binary operation. [definition: Addition on Natural Numbers] Addition on $\mathbb{N}$ is the binary operation \begin{align*} +: \mathbb{N} \times \mathbb{N} &\to \mathbb{N} \end{align*} determined by the recursive clauses \begin{align*} n+1 &= S(n) \end{align*} and \begin{align*} n+S(m) &= S(n+m) \end{align*} for all $m,n \in \mathbb{N}$. [/definition] Because $1$ is the first natural number in this convention, the displayed rules define addition by recursion on the second input. The rule $n+1=S(n)$ says that adding $1$ means taking one successor step, and the rule $n+S(m)=S(n+m)$ says that adding the successor of $m$ means taking one further successor step after forming $n+m$. After addition has encoded repeated successor, the next counting task is repeated grouping. Multiplication records the size of equal groups accumulated by addition, which is why products appear whenever arrays, rectangular grids, finite Cartesian products, or repeated choices are counted. [definition: Multiplication on Natural Numbers] Multiplication on $\mathbb{N}$ is the binary operation \begin{align*} \cdot: \mathbb{N} \times \mathbb{N} &\to \mathbb{N} \end{align*} determined by the recursive clauses \begin{align*} n \cdot 1 &= n \end{align*} and \begin{align*} n \cdot S(m) &= n \cdot m+n \end{align*} for all $m,n \in \mathbb{N}$. [/definition] Many uses of natural numbers ask not how many objects there are, but which count comes earlier. A usable counting system must distinguish being before, being equal, and being after. The natural order formalises that comparison by measuring the positive gap from one count to another. [definition: Natural Order] For $m,n \in \mathbb{N}$, define $m < n$ if there exists $k \in \mathbb{N}$ such that $m+k=n$. Define $m \le n$ if $m<n$ or $m=n$. [/definition] Some constructions need a zero value for empty counts, remainders, and additive identities. Rather than changing the convention for $\mathbb{N}$ itself, it is better to name the enlarged set explicitly so that induction bases and indexing conventions remain unambiguous. [definition: Natural Numbers with Zero] The set of natural numbers with zero is $\mathbb{N}_0 := \{0\} \cup \mathbb{N}$. Its order extends the natural order on $\mathbb{N}$ by $0 \le 0$, by $0<n$ for every $n \in \mathbb{N}$, and by the natural order between elements of $\mathbb{N}$. [/definition] Once the order has been defined, it remains to check that it has the step-by-step shape promised by counting. Ordered sets such as $\mathbb{Q}$ and $\mathbb{R}$ have no immediate next element after a given point, so a theorem is needed to separate the natural order from dense order and to justify arguments that move one successor step at a time. [quotetheorem:9512] This discreteness is what makes induction by a single successor step possible. If $m<n$ in $\mathbb{N}$, there is a first count after $m$, namely $S(m)$, and no natural number can be inserted between them. By contrast, ordered sets such as $\mathbb{Q}$ and $\mathbb{R}$ are dense: between two distinct elements there is always another one. Thus arguments that advance from $n$ to $S(n)$ are special to discrete ordered structures and should not be confused with order arguments in continuous settings. ## Equivalent Characterisations The Peano axioms are structural, while well-ordering is order-theoretic. The two viewpoints answer different questions: Peano axioms explain how to build and reason by successor; well-ordering explains why every nonempty collection of natural numbers has a first element. Ordinary successor induction is enough when the step from $n$ to $S(n)$ only needs the immediately previous case. Many arguments are not so local: to prove a statement at $n$, one may need the statement for every smaller natural number, as in factorisation, divisibility, and recursive algorithms whose next move can jump down by more than one. Strong induction packages this stronger hypothesis while still expressing the same foundational idea that there is no first failure. [quotetheorem:720] Strong induction differs from ordinary successor induction in the information allowed at the inductive step. Instead of assuming only the immediately preceding case, the proof of the case $n$ may use all cases below $n$. This is exactly what is needed when the natural descent in a problem does not land at $n-1$: a divisor may leave a smaller quotient, a recursive algorithm may jump to a much smaller input, and a factorisation argument may split a number into several smaller pieces. The principle is still limited to statements indexed by natural numbers and still requires a genuine argument for every possible next index; it does not replace the work of proving the inductive step. Its main use below is to justify arguments where every attempted counterexample can be reduced to a smaller one. Sometimes a proof needs to know that a counterexample, if it exists, has a least instance. That form is especially natural in divisibility, minimal counterexample arguments, and algorithms that terminate by decreasing a natural number. [quotetheorem:721] At this point there are two powerful reasoning tools on the table: induction builds results by successor steps, while well-ordering chooses the first possible obstruction. To use them interchangeably without changing the foundations, the page needs the precise result that neither principle is stronger than the other. [quotetheorem:722] This equivalence is useful because it lets a proof use whichever form exposes the obstruction most directly without adding a new axiom. Induction is usually cleaner when the property is meant to be built for all natural numbers in order; well-ordering is usually cleaner when the argument starts by assuming a counterexample and then studying the least one. The result also marks a limitation: both methods rely on the special discreteness of $\mathbb{N}$, so they should not be transferred automatically to ordered sets where nonempty subsets may fail to have a least element. Below, this same successor-based structure reappears not as a proof method but as a way to define functions one stage at a time. Recursive definition is another face of the same structure. A function out of $\mathbb{N}$ is determined once its first value and successor rule are determined, which is why natural numbers serve as the basic domain for sequences and iterative processes. [quotetheorem:9513] This theorem justifies definitions such as factorials, powers, recursively defined sequences, and algorithmic iteration. It also explains why the natural numbers act as the universal clock for discrete processes. ## Standard Examples Finite counting is the bridge from natural numbers to cardinality. Every nonempty finite set has a size measured by a natural number, while the empty finite set has cardinality $0$ in $\mathbb{N}_0$. Bijections are the mechanism that makes this measurement independent of labels. [example: Counting a Finite Set] Let $A=\{a,b,c,d\}$, and define $f:\{1,2,3,4\}\to A$ by $f(1)=a$, $f(2)=b$, $f(3)=c$, and $f(4)=d$. To check that $f$ is injective, take $i,j\in\{1,2,3,4\}$ and suppose $f(i)=f(j)$. The four values of $f$ are distinct because $a,b,c,d$ are distinct elements of the set $A$, so equality of outputs forces the same input: \begin{align*} f(i)=f(j)\implies i=j. \end{align*} Thus $f$ is injective. To check that $f$ is surjective, let $x\in A$. Since $A=\{a,b,c,d\}$, exactly one of the following alternatives holds: $x=a$, $x=b$, $x=c$, or $x=d$. In those four cases respectively, \begin{align*} x=f(1) \end{align*} \begin{align*} x=f(2) \end{align*} \begin{align*} x=f(3) \end{align*} \begin{align*} x=f(4). \end{align*} Hence every element of $A$ is hit by $f$, so $f$ is surjective. Therefore $f$ is a bijection from $\{1,2,3,4\}$ to $A$, and the finite set $A$ has cardinality $4$. If the labels are reordered, for example by defining $g(1)=b$, $g(2)=d$, $g(3)=a$, and $g(4)=c$, the same injectivity and surjectivity checks apply: each element of $A$ appears exactly once among the four outputs. The bijection changes, but the counting number measuring the size of $A$ remains $4$. [/example] ## Properties The first major property of $\mathbb{N}$ is that recursive arithmetic behaves like the arithmetic learned from counting. These algebraic laws are the part of semiring arithmetic visible before zero is adjoined: addition and multiplication have the expected commutative, associative, and distributive behaviour, while the additive identity lives in $\mathbb{N}_0$ rather than in $\mathbb{N}$. [remark: Arithmetic Laws for Natural Numbers] For $a,b,c\in\mathbb{N}$, the recursively defined operations satisfy the commutative and associative laws for addition: $a+b=b+a$ and $(a+b)+c=a+(b+c).$ They also satisfy the commutative and associative laws for multiplication: $a\cdot b=b\cdot a$ and $(a\cdot b)\cdot c=a\cdot (b\cdot c).$ Finally, multiplication distributes over addition: $a\cdot(b+c)=a\cdot b+a\cdot c.$ Here multiplication is the operation determined by the clauses $a\cdot 1=a$ and $a\cdot S(b)=a\cdot b+a$. [/remark] These laws are not extra assumptions about familiar arithmetic; they are consequences of the recursive definitions of addition and multiplication. Associativity lets repeated sums be written without tracking parentheses, commutativity says that counting two finite groups does not depend on their order, and distributivity explains why multiplication behaves as repeated addition. Many later constructions depend on these facts: products of finite sets, powers, divisibility, and algebraic simplification all require the recursively defined operations to obey these stable laws. Inequalities must also survive the basic operations if the order on $\mathbb{N}$ is to support estimates, divisibility comparisons, and ordered algebra. The next property records that adding the same amount, or multiplying by the same positive amount, preserves strict order. [quotetheorem:9515] The positivity condition is built into $\mathbb{N}$ under the convention used on this page, so multiplying a strict inequality by a natural number cannot collapse both sides to the same value. This is why the corresponding statement over $\mathbb{N}_0$ would need to exclude multiplication by $0$: from $a<b$ one cannot conclude $0\cdot a<0\cdot b$. The theorem is used whenever an inequality is rescaled or translated, for example when comparing multiples in divisibility arguments, bounding finite sums, or proving that an iterative process remains ordered after the same operation is applied at each step. Passing to $\mathbb{N}_0$ adds an additive identity, while passing to the integers, denoted $\mathbb{Z}$, adds additive inverses. Here $\mathbb{Z}$ is the usual integer system containing $\mathbb{N}_0$. Arithmetic on $\mathbb{N}_0$ uses the usual extension of the recursive operations on $\mathbb{N}$: $0+a=a+0=a$ and $0\cdot a=a\cdot 0=0$ for every $a \in \mathbb{N}_0$, while addition and multiplication between positive natural numbers are the operations already defined. Divisibility requires a way to compare an arbitrary natural number with multiples of another one. The key fact is that after subtracting enough copies of $b$, the leftover part can be chosen smaller than $b$ and is uniquely determined. The quotient and remainder naturally live in $\mathbb{N}_0$, using the extended order just defined. [quotetheorem:9516] The remainder condition $0\le r<b$ is what makes the representation useful rather than merely possible. Without the upper bound on $r$, one could trade a copy of $b$ between the quotient and the remainder in many ways; for example, $17=3\cdot 5+2$ also equals $2\cdot 5+7$, but only the first expression has remainder smaller than $5$. Uniqueness is equally important: it lets “the quotient” and “the remainder” be treated as well-defined outputs of division, which is the basis for modular arithmetic and for the Euclidean algorithm. The theorem applies to division by a positive natural number $b$; it does not allow division by $0$, and the use of $\mathbb{N}_0$ is necessary because either the quotient or the remainder may be $0$. Termination proofs need a measure that cannot decrease forever. Natural numbers supply such a measure because every nonempty collection has a least element. The descent principle packages this idea for algorithms, minimal-counterexample arguments, and iterative reductions. [quotetheorem:4522] This principle is the hidden engine behind the Euclidean algorithm: each nonzero remainder is a smaller natural number than the previous divisor. It is also the formal reason that minimal-counterexample proofs cannot descend forever. ## Relationship to Other Concepts Natural numbers sit at the base of several number systems. The integers extend them by adding $0$ and negatives; the rational numbers arise from ratios of integers; the [real numbers](/page/Real%20Numbers) complete the rationals for limits and analysis. Each extension solves a limitation of the previous system while preserving the natural numbers inside it. The set $\mathbb{N}$ is also the standard indexing set for a [sequence](/page/Sequence). A sequence in a set $X$ is a [function](/page/Function) $a: \mathbb{N} \to X$, and the notation $a_n$ abbreviates $a(n)$. This is why induction and recursion are so closely connected to convergence, recurrence relations, and discrete dynamical systems. In [set theory](/page/Set), the natural numbers provide the first infinite set and the reference model for countability. A set is countably infinite when it can be placed in bijection with $\mathbb{N}$, and finite sets are measured by initial segments of $\mathbb{N}$. In number theory, the natural numbers are the basic objects of divisibility, prime factorisation, congruences, arithmetic functions, and [Diophantine equations](/page/Diophantine%20Equations). The fact that every nonempty subset has a least element is one of the main reasons elementary arguments in number theory can be so powerful. The convention $\mathbb{N}=\{1,2,3,\ldots\}$ should always be kept in view. When a statement needs zero, use $\mathbb{N}_0$ explicitly. This prevents ambiguity in formulas involving induction bases, sequence indexing, and probability distributions such as the geometric distribution. ## References Peano, *Arithmetices principia, nova methodo exposita* (1889). Dedekind, *Was sind und was sollen die Zahlen?* (1888). Enderton, *Elements of Set Theory* (1977). Hardy and Wright, *An Introduction to the Theory of Numbers* (1938).