A [polynomial ring](/page/Polynomial%20Ring) over a field looks enormous: it contains polynomials of every degree, ideals can be generated by many equations, and algebraic sets can be cut out by apparently infinite systems. The surprise behind Noetherian rings is that, in many rings that arise in algebra, this infinitude is illusory. If the ring is Noetherian, every ideal has a finite list of generators, every ascending process of enlarging ideals eventually stops, and many constructions become finite enough to compute with.
The opposite behaviour is not a technical nuisance; it changes the subject. In a non-Noetherian ring, an ideal may require infinitely many independent generators, so an algebraic condition cannot be compressed into finitely many equations. Noetherian rings isolate the rings in which ideal-theoretic information is finite in a strong structural sense.
Throughout the main ring-theoretic sections of this chapter, rings are commutative and have a multiplicative identity unless a statement explicitly says otherwise. The later module section briefly records the left-module version because it is the form needed in broader algebra, but the central page topic is the commutative Noetherian ring.
[example: An Ideal That Never Becomes Finitely Generated]
Let $k$ be a field and let
\begin{align*}
R = k[x_1,x_2,x_3,\ldots]
\end{align*}
be the polynomial ring in countably many variables. Consider the ideal
\begin{align*}
I = (x_1,x_2,x_3,\ldots) \trianglelefteq R.
\end{align*}
We show that $I$ cannot be generated by finitely many elements.
Suppose, for contradiction, that
\begin{align*}
I = (f_1,\ldots,f_m)
\end{align*}
for some $f_1,\ldots,f_m \in R$. Each polynomial is a finite $k$-linear combination of monomials, and each monomial uses only finitely many variables, so the finite list $f_1,\ldots,f_m$ involves only finitely many variables altogether. Hence there exists $N \in \mathbb{N}$ such that
\begin{align*}
f_i \in k[x_1,\ldots,x_N] \quad \text{for every } 1 \le i \le m.
\end{align*}
Since $f_i \in I$, no $f_i$ has a nonzero constant term: indeed, every element of $I=(x_1,x_2,\ldots)$ is a finite sum of terms $a_jx_j$, and each term $a_jx_j$ has zero constant term. Therefore, when $f_i$ is regarded as a polynomial in $k[x_1,\ldots,x_N]$, every monomial appearing in $f_i$ is divisible by at least one of $x_1,\ldots,x_N$. Thus
\begin{align*}
f_i \in (x_1,\ldots,x_N)R \quad \text{for every } 1 \le i \le m.
\end{align*}
Because $(x_1,\ldots,x_N)R$ is an ideal, every $R$-linear combination of the $f_i$ also lies in $(x_1,\ldots,x_N)R$. Hence
\begin{align*}
(f_1,\ldots,f_m) \subseteq (x_1,\ldots,x_N)R.
\end{align*}
Using the assumed equality $I=(f_1,\ldots,f_m)$, this gives
\begin{align*}
I \subseteq (x_1,\ldots,x_N)R.
\end{align*}
But $x_{N+1} \in I$ by the definition of $I$. To see that $x_{N+1} \notin (x_1,\ldots,x_N)R$, define the $k$-algebra homomorphism
\begin{align*}
\varphi:R \to k[x_{N+1},x_{N+2},\ldots]
\end{align*}
by $\varphi(x_j)=0$ for $1 \le j \le N$ and $\varphi(x_j)=x_j$ for $j>N$. Every generator $x_1,\ldots,x_N$ of $(x_1,\ldots,x_N)R$ maps to $0$, so every element of $(x_1,\ldots,x_N)R$ maps to $0$. However,
\begin{align*}
\varphi(x_{N+1}) = x_{N+1} \ne 0.
\end{align*}
Thus $x_{N+1} \notin (x_1,\ldots,x_N)R$, contradicting $I \subseteq (x_1,\ldots,x_N)R$. Therefore $I$ is not finitely generated.
[/example]
This example shows the failure that Noetherianity forbids. It is not about the cardinality of the ring: even countable rings can be Noetherian or non-Noetherian. The issue is whether ideals can be controlled by finite data.
## Definition
Before naming auxiliary chain language, recall the parent object. A Noetherian ring is not a new kind of algebraic operation; it is a finiteness condition imposed on a commutative [ring](/page/Ring). The first problem is to make sense of an ideal-building process that never runs out of new information. The definition records exactly the demand that every such process must stop.
[definition: Noetherian Ring]
A commutative ring $R$ is a Noetherian ring if for every sequence of ideals
\begin{align*}
I_1 \subset I_2 \subset I_3 \subset \cdots
\end{align*}
there exists $N \in \mathbb{N}$ such that $I_n = I_N$ for all $n \ge N$.
[/definition]
This definition turns the informal phrase "eventually no new ideal appears" into a precise stopping rule. To compare Noetherianity with other finiteness conditions, it helps to separate the stopping mechanism from the ring name itself. The next definition names the chain condition directly, so later arguments can point to the exact obstruction: an infinite process of enlarging ideals that never stabilises.
[definition: Ascending Chain Condition on Ideals]
Let $R$ be a commutative ring. The ring $R$ satisfies the ascending chain condition on ideals if for every sequence of ideals
\begin{align*}
I_1 \subset I_2 \subset I_3 \subset \cdots
\end{align*}
there exists $N \in \mathbb{N}$ such that $I_n = I_N$ for all $n \ge N$.
[/definition]
Thus a commutative ring is Noetherian exactly when it satisfies the ascending chain condition on ideals. The definition is stated using chains, but in practice one usually checks finite generation of ideals. This matters because a finite generating set is explicit algebraic data, while a universal statement about chains is often hard to inspect directly. The next theorem is the basic equivalence that makes the concept usable.
[quotetheorem:9999]
This equivalence is the working form of the definition. It says that Noetherianity is not merely a condition on chains; it is exactly the assertion that every ideal admits a finite description.
A second formulation is useful when one wants to extract a maximal ideal with some property. Instead of building an infinite chain, one asks whether every nonempty collection of ideals has a maximal member under inclusion.
[quotetheorem:10000]
The maximal condition is a compactness principle for ideals. It lets arguments choose a largest counterexample, then enlarge it to get a contradiction.
The smallest examples come from rings where ideals are already well understood. Principal ideal domains provide the first reliable supply because their ideals are generated by a single element.
[example: Integers as a Noetherian Ring]
Let $I \trianglelefteq \mathbb{Z}$ be an ideal. If $I=\{0\}$, then $I=(0)$, so $I$ is generated by one element. If $I\ne \{0\}$, choose the least positive integer $n\in I$; this exists by the [well-ordering principle](/theorems/721), since $I$ contains some nonzero integer and therefore contains a positive one.
We show that $I=(n)$. Because $n\in I$ and $I$ is closed under multiplication by arbitrary integers, every multiple $qn$ lies in $I$, so $(n)\subseteq I$. Conversely, let $a\in I$. By Euclidean division, there exist $q,r\in\mathbb{Z}$ such that
\begin{align*}
a=qn+r \quad \text{with} \quad 0\le r<n.
\end{align*}
Since $a\in I$ and $qn\in I$, closure of $I$ under subtraction gives
\begin{align*}
r=a-qn\in I.
\end{align*}
If $r>0$, then $r$ is a positive element of $I$ smaller than $n$, contradicting the choice of $n$. Hence $r=0$, so $a=qn\in(n)$. Thus $I\subseteq(n)$, and therefore $I=(n)$.
Every ideal of $\mathbb{Z}$ is therefore generated by one element, so by *[Ideal Characterisation of Noetherian Rings](/theorems/9999)*, $\mathbb{Z}$ is Noetherian. The same reasoning applies to any [principal ideal domain](/page/Principal%20Ideal%20Domain), because in such a ring every ideal is principal by definition.
[/example]
A field is even smaller from the ideal-theoretic viewpoint. Its only ideals are $0$ and the whole field, so no infinite growth of ideals can begin.
[example: Fields and Polynomial Rings in One Variable]
Let $k$ be a field. We first identify its ideals. If $I \trianglelefteq k$ and $I=\{0\}$, then $I=(0)$. If $I\ne \{0\}$, choose $a\in I$ with $a\ne 0$. Since $k$ is a field, $a^{-1}\in k$, and closure of the ideal $I$ under multiplication by elements of $k$ gives
\begin{align*}
1=a^{-1}a\in I.
\end{align*}
For any $b\in k$, closure under multiplication by $b$ gives
\begin{align*}
b=b\cdot 1\in I.
\end{align*}
Thus $I=k$. Hence the only ideals of $k$ are $(0)$ and $(1)=k$, so every ideal of $k$ is finitely generated. By *Ideal Characterisation of Noetherian Rings*, $k$ is Noetherian.
Now consider $k[x]$. Since $k[x]$ is a principal ideal domain, every ideal $J\trianglelefteq k[x]$ has the form
\begin{align*}
J=(f)
\end{align*}
for some $f\in k[x]$. This means every element of $J$ is of the form $g(x)f$ with $g(x)\in k[x]$, so $J$ is generated by the single element $f$. Therefore every ideal of $k[x]$ is finitely generated, and by *Ideal Characterisation of Noetherian Rings*, $k[x]$ is Noetherian.
[/example]
These examples foreshadow [Hilbert's basis theorem](/theorems/2904): starting with a Noetherian ring, adjoining finitely many polynomial variables preserves Noetherianity. That theorem is the reason Noetherian rings dominate affine algebraic geometry.
## Chains, Generators, and Maximal Counterexamples
### Stabilisation
The three faces of Noetherianity are chain stopping, finite generation, and maximal elements. Each viewpoint solves a different problem. Chains measure processes, generators measure data, and maximal elements power existence arguments.
A chain of ideals is not merely a sequence; it is an algebraic process in which more relations, more generators, or more equations are added stage by stage. To use the chain condition in arguments, it helps to name the moment when the process has stopped changing.
[definition: Stabilising Chain of Ideals]
Let $R$ be a commutative ring. An ascending chain of ideals
\begin{align*}
I_1 \subset I_2 \subset I_3 \subset \cdots
\end{align*}
stabilises if there exists $N \in \mathbb{N}$ such that $I_n = I_N$ for every $n \ge N$.
[/definition]
Stabilisation is stronger than saying that the union has a finite generating set in the abstract; it says the finite information already appears at a finite stage of the process.
[example: A Non-Stabilising Chain]
Let $R = k[x_1,x_2,x_3,\ldots]$, and for each $n \ge 1$ set
\begin{align*}
I_n=(x_1,\ldots,x_n).
\end{align*}
Since every generator $x_1,\ldots,x_n$ of $I_n$ is also among the generators of $I_{n+1}=(x_1,\ldots,x_n,x_{n+1})$, every $R$-linear combination of $x_1,\ldots,x_n$ lies in $I_{n+1}$. Thus
\begin{align*}
I_n \subseteq I_{n+1}.
\end{align*}
Therefore the ideals form an ascending chain
\begin{align*}
I_1 \subseteq I_2 \subseteq I_3 \subseteq \cdots .
\end{align*}
The inclusion is strict for every $n$. We have $x_{n+1}\in I_{n+1}$ because $x_{n+1}$ is one of the generators of $I_{n+1}$. To show that $x_{n+1}\notin I_n$, define the $k$-algebra homomorphism
\begin{align*}
\varphi:R\to k[x_{n+1},x_{n+2},\ldots]
\end{align*}
by $\varphi(x_j)=0$ for $1\le j\le n$ and $\varphi(x_j)=x_j$ for $j>n$. If $f\in I_n$, then for some $a_1,\ldots,a_n\in R$,
\begin{align*}
f=a_1x_1+\cdots+a_nx_n.
\end{align*}
Applying $\varphi$ gives
\begin{align*}
\varphi(f)=\varphi(a_1)\varphi(x_1)+\cdots+\varphi(a_n)\varphi(x_n)=\varphi(a_1)0+\cdots+\varphi(a_n)0=0.
\end{align*}
But
\begin{align*}
\varphi(x_{n+1})=x_{n+1}\ne 0.
\end{align*}
Hence $x_{n+1}\notin I_n$, so
\begin{align*}
I_n \subsetneq I_{n+1}
\end{align*}
for every $n\ge 1$.
Thus the chain never reaches an index $N$ after which all ideals are equal. It does not stabilise, so $R$ fails the ascending chain condition on ideals and is not Noetherian.
[/example]
### Principal Ideals
The chain viewpoint gives a direct way to detect failure. The finite-generation viewpoint gives a direct way to prove success, especially when ideals have a normal form. Principal ideal domains are the cleanest case: every ideal is described by one element, so the Noetherian condition is already present in the ideal structure.
[remark: Principal Ideals Give Noetherianity]
A principal ideal domain is an [integral domain](/page/Integral%20Domain) in which every ideal has the form $(a)$ for some element $a$ of the ring. Since each ideal is generated by a single element, every ideal is finitely generated. Thus every principal ideal domain is Noetherian.
[/remark]
This explains why Euclidean domains, such as $\mathbb{Z}$ and $k[x]$ for a field $k$, satisfy Noetherianity. Their ideals are generated by one element, so the full finite-generation condition is built into the definition of a principal ideal domain.
### Noetherian Induction
A typical proof by maximal counterexample begins with a collection of ideals that fail a desired property. In a Noetherian ring, such a collection has a maximal member, so there is a largest obstruction to eliminate. This transforms an infinite search into a single obstruction that can be attacked.
[example: How a Maximal Counterexample Argument Starts]
Let $R$ be a Noetherian ring, and let $\mathcal{S}$ be a nonempty collection of ideals of $R$ whose elements are the obstructions under consideration. By the maximal-condition characterisation of Noetherian rings, the collection $\mathcal{S}$ has a maximal element under inclusion. Thus there exists an ideal $I\in \mathcal{S}$ such that for every $K\in \mathcal{S}$,
\begin{align*}
I\subseteq K \quad \text{implies} \quad K=I.
\end{align*}
A typical maximal-counterexample argument now tries to contradict this property. Suppose the argument constructs an ideal $J\trianglelefteq R$ with
\begin{align*}
I\subsetneq J
\end{align*}
and also proves that
\begin{align*}
J\in \mathcal{S}.
\end{align*}
Since $I\subsetneq J$, we have $I\subseteq J$ and $J\ne I$. But $J\in\mathcal{S}$, so the defining property of the chosen maximal element $I$ says that $I\subseteq J$ must imply $J=I$. This contradicts $J\ne I$.
Therefore, in a Noetherian ring, once a maximal obstruction $I$ has been chosen, any construction that produces a strictly larger obstruction $J$ eliminates the possibility that $\mathcal{S}$ contained obstructions in the first place.
[/example]
The example is schematic, but the pattern is pervasive. It is the ideal-theoretic analogue of ordinary induction on $\mathbb{N}$: instead of reducing an integer, one enlarges an ideal until the Noetherian condition forbids further movement.
## Polynomial Rings and Finite Algebraic Data
The most important permanence property of Noetherian rings is that adjoining finitely many polynomial variables preserves Noetherianity. This is where the condition becomes a theorem about equations: a polynomial ideal in finitely many variables over a field has finite generators.
Before stating the theorem, it is worth naming the construction it controls. Polynomial rings can create many new ideals, so preservation of Noetherianity is not automatic from the definition.
[definition: Polynomial Ring over a Ring]
Let $R$ be a commutative ring. The polynomial ring $R[x]$ is the ring whose elements are finite sums
\begin{align*}
f = a_0 + a_1x + \cdots + a_nx^n
\end{align*}
with $a_i \in R$, equipped with the usual addition and multiplication of polynomials.
[/definition]
The word "finite" in the definition of a polynomial is not enough to make all ideals finite. An ideal in $R[x]$ may contain infinitely many polynomials of unrelated degrees, and Hilbert's theorem says that Noetherianity still prevents infinite independent generation.
[quotetheorem:860]
[Hilbert's basis theorem](/theorems/2907) is the engine behind finite generation in commutative algebra. The theorem applies one variable at a time, and this is exactly what is needed for the finite-variable polynomial rings that occur in algebraic geometry and elimination theory. Iteration turns a one-variable permanence theorem into the standard finite-dimensional polynomial case.
[quotetheorem:10001]
This theorem is the algebraic reason that affine varieties can be described by finite systems of polynomial equations, even when they are initially presented by infinitely many equations.
[example: Infinite Equations Collapse to Finitely Many]
Let $k$ be a field, and write $\mathbb{A}^2_k$ for the affine plane over $k$, whose points are pairs $(a,b)$ with $a,b\in k$. Let $X\subset \mathbb{A}^2_k$, and let $I\trianglelefteq k[x,y]$ be the ideal generated by the polynomials that vanish on every point of $X$. Since $k$ is Noetherian and polynomial rings in finitely many variables over a Noetherian ring are Noetherian by *[Hilbert Basis Theorem for Finite Polynomial Rings](/theorems/10001)*, the ring $k[x,y]$ is Noetherian. By *Ideal Characterisation of Noetherian Rings*, the ideal $I$ has a finite generating set, so there are $f_1,\ldots,f_m\in I$ with
\begin{align*}
I=(f_1,\ldots,f_m).
\end{align*}
We now check that the zero locus of all elements of $I$ is exactly the zero locus of this finite family. If $p=(a,b)\in \mathbb{A}^2_k$ satisfies $g(p)=0$ for every $g\in I$, then in particular $f_i(p)=0$ for every $1\le i\le m$, because each $f_i$ lies in $I$.
Conversely, suppose $p=(a,b)$ satisfies
\begin{align*}
f_1(p)=\cdots=f_m(p)=0.
\end{align*}
Let $g\in I$. Since $I=(f_1,\ldots,f_m)$, there exist $h_1,\ldots,h_m\in k[x,y]$ such that
\begin{align*}
g=h_1f_1+\cdots+h_mf_m.
\end{align*}
Evaluating at $p$ gives
\begin{align*}
g(p)=h_1(p)f_1(p)+\cdots+h_m(p)f_m(p).
\end{align*}
Using $f_i(p)=0$ for each $i$, this becomes
\begin{align*}
g(p)=h_1(p)0+\cdots+h_m(p)0=0.
\end{align*}
Thus every element of $I$ vanishes at $p$.
Therefore the common zero locus cut out by all elements of $I$ is the same as the common zero locus cut out by the finite family $f_1,\ldots,f_m$. Noetherianity turns the possibly infinite list of equations defining $I$ into finitely many equations without changing the resulting algebraic set.
[/example]
This finite collapse is not available in infinitely many variables. The earlier chain $(x_1) \subset (x_1,x_2) \subset \cdots$ records exactly the obstruction.
## Quotients, Localisation, and Finite Algebras
### Quotients
Noetherianity would be much less useful if it were destroyed by ordinary algebraic constructions. Fortunately, it behaves well under the operations used to pass to coordinate rings, impose relations, and study local behaviour.
A [quotient ring](/page/Quotient%20Ring) is how algebra enforces equations. If one passes from $R$ to $R/I$, every ideal of the quotient corresponds to an ideal of $R$ containing $I$, so the Noetherian condition descends. The construction deserves to be stated because it is the bridge from polynomial rings to coordinate rings.
[definition: Quotient Ring by an Ideal]
Let $R$ be a ring and let $I \trianglelefteq R$ be an ideal. The quotient ring $R/I$ is the ring of additive cosets $r + I$, with addition and multiplication induced by the operations of $R$.
[/definition]
Passing to $R/I$ can collapse distinct elements and replace ideals by cosets, so a priori finite generation of ideals might be lost when relations are imposed. The key point is that ideals of $R/I$ are controlled by ideals of $R$ containing $I$, which means the ascending-chain obstruction cannot reappear after quotienting a Noetherian ring.
[remark: Quotients Preserve Noetherianity]
If $R$ is Noetherian and $I \trianglelefteq R$ is an ideal, then the quotient ring $R/I$ is Noetherian. Indeed, every ideal of $R/I$ corresponds to an ideal of $R$ that contains $I$, so an ascending chain of ideals in $R/I$ lifts to an ascending chain in $R$ and must stabilise.
[/remark]
This preservation property is why finitely generated $k$-algebras of the form $k[x_1, \ldots, x_n]/I$ are Noetherian. The polynomial ring is Noetherian by Hilbert's theorem, and the quotient preserves the property.
### Localisation
Localisation is subtler than taking a quotient. It introduces denominators, so ideals can change shape, and the ring may acquire new units. To choose the denominators coherently, one first singles out a set closed under multiplication.
[definition: Multiplicative Set]
Let $R$ be a commutative ring. A subset $S \subset R$ is a multiplicative set if $1_R \in S$ and $st \in S$ for all $s,t \in S$.
[/definition]
A multiplicative set only records which elements should become invertible; it does not yet say when two fractions should count as the same. The obstruction is that denominators may include zero divisors, so equality of fractions must allow multiplication by another element of $S$ before the usual cross-multiplication test is valid.
[definition: Localisation of a Ring]
Let $R$ be a commutative ring and let $S \subset R$ be a multiplicative set. The localisation $S^{-1}R$ is the ring whose elements are equivalence classes of pairs $(r,s) \in R \times S$, written $r/s$, where $(r,s) \sim (r',s')$ if there exists $t \in S$ such that
\begin{align*}
t(s'r - sr') = 0.
\end{align*}
Addition and multiplication are defined by
\begin{align*}
\frac{r}{s} + \frac{r'}{s'} &= \frac{rs' + r's}{ss'}, & \frac{r}{s}\frac{r'}{s'} &= \frac{rr'}{ss'}.
\end{align*}
[/definition]
Introducing denominators can make many elements newly invertible and can identify ideals that looked different in the original ring, so it is not automatic that finite ideal control survives. The next structural question is whether localisation respects the same finite ideal condition as quotienting: after passing to $S^{-1}R$, every ideal should still be controlled by finitely many generators coming from the original Noetherian ring. This permanence result is what allows local arguments to use denominators without leaving the Noetherian world.
[quotetheorem:10002]
Because both quotients and localisations preserve Noetherianity, the property is well adapted to algebraic geometry. Coordinate rings, local rings at points, and rings of regular functions on affine pieces remain inside the same finite framework.
[example: Local Rings of Affine Varieties]
In this example, a prime ideal means an ideal $\mathfrak{p}\ne A$ such that $ab\in\mathfrak{p}$ forces $a\in\mathfrak{p}$ or $b\in\mathfrak{p}$. For a ring $A$ and a prime ideal $\mathfrak{p}$, the notation $A_{\mathfrak{p}}$ denotes the localization obtained by allowing denominators from $A\setminus\mathfrak{p}$.
Let $k$ be a field, let
\begin{align*}
A = k[x_1,\ldots,x_n]/I
\end{align*}
for an ideal $I \trianglelefteq k[x_1,\ldots,x_n]$, and let $\mathfrak{p} \trianglelefteq A$ be a prime ideal. Since $k$ is a field, its only ideals are $(0)$ and $(1)$, so every ideal of $k$ is finitely generated. Hence $k$ is Noetherian by *Ideal Characterisation of Noetherian Rings*. Applying *Finite Polynomial Rings over Noetherian Rings* gives that
\begin{align*}
k[x_1,\ldots,x_n]
\end{align*}
is Noetherian. Since $A$ is the quotient of this Noetherian ring by the ideal $I$, the quotient permanence result above gives that $A$ is Noetherian.
Now set
\begin{align*}
S=A\setminus \mathfrak{p}.
\end{align*}
Because $\mathfrak{p}$ is prime, $1\notin \mathfrak{p}$, so $1\in S$. If $s,t\in S$ and $st\notin S$, then $st\in\mathfrak{p}$; primeness would imply $s\in\mathfrak{p}$ or $t\in\mathfrak{p}$, contradicting $s,t\in S$. Thus $st\in S$, and $S$ is a multiplicative set. The local ring at $\mathfrak{p}$ is therefore the localisation
\begin{align*}
A_{\mathfrak{p}}=(A\setminus\mathfrak{p})^{-1}A=S^{-1}A.
\end{align*}
Since $A$ is Noetherian and $S$ is multiplicative, the localisation permanence result above implies that $A_{\mathfrak{p}}$ is Noetherian.
Thus passing from an affine [coordinate ring](/page/Coordinate%20Ring) to the local ring at a prime ideal preserves the finite generation of ideals.
[/example]
This is the algebraic reason that local calculations on affine varieties still involve finitely generated ideals. Passing to a point does not lose the finite nature of the coordinate ring.
## Noetherian Modules
### Submodules
Rings are often studied through their modules. The module version of Noetherianity is more flexible: a ring is Noetherian exactly when it is Noetherian as a module over itself, and many arguments about ideals are module arguments in disguise.
The module definition mirrors the ring definition but replaces ideals by submodules. This matters because kernels, images, and presentations naturally live in module categories, and a finiteness condition that works only for ideals would be too narrow for homological algebra.
[definition: Noetherian Module]
Let $R$ be a ring and let $M$ be a left $R$-module. The module $M$ is Noetherian if every ascending chain of submodules
\begin{align*}
M_1 \subset M_2 \subset M_3 \subset \cdots
\end{align*}
stabilises.
[/definition]
A module may satisfy a stopping condition on chains, but computations with kernels, relations, and presentations require explicit finite generating sets for submodules. The bridge is that, under the Noetherian condition, the absence of endlessly growing submodules is equivalent to finite control over every submodule.
[quotetheorem:2815]
In the commutative setting of this page, this theorem recovers the ring statement by taking $M = R$ as a module over itself. Its submodules are precisely the ideals of $R$. In noncommutative algebra one must distinguish left Noetherian, right Noetherian, and two-sided ideal conditions.
### Finite Modules
Finite direct sums preserve Noetherianity, so finite free modules over a Noetherian ring behave like finite-dimensional vector spaces in the specific sense that their submodules remain finitely generated. In practice, however, finitely generated modules are rarely presented only as free modules: they occur as quotients of free modules, as images of maps, and as modules with relations imposed. The key question is therefore whether Noetherian control descends from a finite free presentation to every [finitely generated module](/page/Finitely%20Generated%20Module) over the ring.
[quotetheorem:2814]
This result is one of the main practical payoffs. It says that once the ring is Noetherian, finite presentations and finite generating sets stay meaningful beyond the free case: arbitrary finitely generated modules, and hence their submodules, remain governed by finite data.
[example: Submodules of $R^n$]
Let $R$ be a commutative Noetherian ring, and let $N \subset R^n$ be an $R$-submodule. Write $e_i$ for the vector whose $i$th coordinate is $1$ and whose other coordinates are $0$. For every vector $(a_1,\ldots,a_n)\in R^n$, we have
\begin{align*}
(a_1,\ldots,a_n)=a_1e_1+\cdots+a_ne_n.
\end{align*}
Thus $R^n$ is generated by the finite set $\{e_1,\ldots,e_n\}$ as an $R$-module.
Since $R$ is Noetherian and $R^n$ is finitely generated, the quoted result on [modules over Noetherian rings](/theorems/2814) implies that $R^n$ is a Noetherian $R$-module. By the submodule characterisation of Noetherian modules, every submodule of a Noetherian module is finitely generated. Applying this to the submodule $N\subset R^n$, there exist vectors $v_1,\ldots,v_m\in N$ such that
\begin{align*}
N=Rv_1+\cdots+Rv_m.
\end{align*}
In particular, if $N$ is the solution module of a homogeneous linear system over $R$, then all solutions are obtained from finitely many solution vectors by taking $R$-linear combinations.
[/example]
### Exact Sequences
The module perspective also explains why exact sequences preserve Noetherian behaviour. An exact sequence divides a module into a submodule and a quotient, so the relevant question is whether finiteness of those two pieces is equivalent to finiteness of the whole module. The answer is yes, and it is one of the standard tools for building Noetherian modules.
[quotetheorem:2811]
This exact-sequence criterion is the algebraic mechanism behind many permanence results. It lets one build Noetherian modules from simpler pieces and test Noetherianity through kernels and quotients.
## Prime Ideals, Decomposition, and Dimension
### Prime Ideals
Noetherian rings do not make prime ideals disappear; instead they make the prime spectrum manageable enough to support decomposition and dimension theory. In commutative algebra, prime ideals are the algebraic shadows of irreducible geometric pieces.
Prime ideals detect when a quotient ring has no zero-divisor decomposition. They are the points of affine schemes and the basic objects of local algebra, so they must be named before discussing decomposition.
[definition: Prime Ideal]
Let $R$ be a commutative ring. An ideal $\mathfrak{p} \trianglelefteq R$ is a prime ideal if $\mathfrak{p} \ne R$ and whenever $ab \in \mathfrak{p}$ for $a,b \in R$, either $a \in \mathfrak{p}$ or $b \in \mathfrak{p}$.
[/definition]
Prime ideals allow a ring to be studied through domains $R/\mathfrak{p}$. For an arbitrary ideal $I$, the first structural question is whether there are prime ideals sitting minimally above it, since those primes describe the irreducible pieces forced by $I$. Existence of such minimal primes is not the special Noetherian miracle; the Noetherian hypothesis becomes important because it leads to finite control over the primes and to decomposition theorems.
[quotetheorem:10003]
Minimal primes identify the irreducible pieces that lie over a given ideal. In the Noetherian setting they become the entry point to a finite decomposition theory, rather than merely an existence statement about primes above $I$.
### Nilpotents and Primary Decomposition
Decomposition theory has to handle nilpotent elements as well as prime ideals. Nilpotents are invisible at the level of ordinary point sets but visible in the ring, so the nilradical measures the part of the ring made from elements that eventually vanish under powers.
[definition: Nilradical]
Let $R$ be a commutative ring. The nilradical of $R$ is
\begin{align*}
\operatorname{nil}(R) = \{a \in R : a^n = 0 \text{ for some } n \in \mathbb{N}\}.
\end{align*}
[/definition]
The nilradical identifies elements that vanish after taking powers. For a general ideal $I \subset R$, the notation $\sqrt{I}$ denotes the radical
\begin{align*}
\sqrt{I}=\{a\in R : a^n\in I \text{ for some } n\ge 1\}.
\end{align*}
The set $\operatorname{Spec}(R)$ consists of the prime ideals of $R$, and $\operatorname{mSpec}(R)$ consists of the maximal ideals. A proper ideal $Q\subset R$ is primary if whenever $ab\in Q$ and $a\notin Q$, then $b^n\in Q$ for some $n\ge 1$; equivalently, every zero divisor in $R/Q$ is nilpotent. A primary decomposition of an ideal $I$ is an expression of $I$ as a finite intersection of primary ideals, often with the prime ideals $\sqrt{Q_i}$ recording the associated prime-like pieces.
With this language in place, the main point is that arbitrary ideals can have several overlapping prime-like components and embedded nilpotent behaviour. The theorem below answers the termination question: under a Noetherian hypothesis, the decomposition into primary pieces can be made finite rather than continuing indefinitely.
[quotetheorem:2885]
Primary decomposition is the ideal-theoretic analogue of factoring an integer into prime powers. It is one of the major reasons Noetherian hypotheses appear in algebraic geometry.
### Dimension
Dimension theory studies chains of prime ideals. Noetherianity does not force all such chains to be short, but it makes dimension theory usable in a broad class of rings. The invariant that records the longest possible prime-chain length is Krull dimension.
[definition: Krull Dimension]
Let $R$ be a commutative ring. The Krull dimension of $R$ is the supremum of all integers $d \ge 0$ for which there exists a strictly increasing chain of prime ideals
\begin{align*}
\mathfrak{p}_0 \subsetneq \mathfrak{p}_1 \subsetneq \cdots \subsetneq \mathfrak{p}_d.
\end{align*}
[/definition]
Krull dimension measures the length of prime-ideal specialisation chains. It is independent of the ACC on all ideals; a Noetherian ring can have positive dimension because prime chains of finite length still occur.
[example: Dimension of a Polynomial Ring over a Field]
Let
\begin{align*}
R=k[x_1,\ldots,x_n].
\end{align*}
For each $0\le r\le n$, set $\mathfrak{p}_r=(x_1,\ldots,x_r)$, with $\mathfrak{p}_0=(0)$. The quotient is
\begin{align*}
R/\mathfrak{p}_r \cong k[x_{r+1},\ldots,x_n],
\end{align*}
where for $r=n$ the right-hand side is $k$. Since a polynomial ring over [a field is an integral domain](/theorems/8301), each quotient $R/\mathfrak{p}_r$ is an integral domain, so each $\mathfrak{p}_r$ is prime by the quotient characterisation of prime ideals.
The inclusions are strict. For $0\le r<n$, the element $x_{r+1}$ belongs to
\begin{align*}
\mathfrak{p}_{r+1}=(x_1,\ldots,x_r,x_{r+1}),
\end{align*}
but its image in
\begin{align*}
R/\mathfrak{p}_r \cong k[x_{r+1},\ldots,x_n]
\end{align*}
is still the nonzero polynomial $x_{r+1}$, so $x_{r+1}\notin \mathfrak{p}_r$. Hence
\begin{align*}
(0)\subsetneq (x_1)\subsetneq (x_1,x_2)\subsetneq \cdots \subsetneq (x_1,\ldots,x_n)
\end{align*}
is a strictly increasing chain of prime ideals of length $n$. By the definition of Krull dimension, this proves
\begin{align*}
\dim k[x_1,\ldots,x_n]\ge n.
\end{align*}
The standard [dimension theorem](/theorems/915) for polynomial rings over fields gives the reverse inequality
\begin{align*}
\dim k[x_1,\ldots,x_n]\le n.
\end{align*}
Combining the two inequalities gives
\begin{align*}
\dim k[x_1,\ldots,x_n]=n.
\end{align*}
Thus the displayed chain is not merely evidence of positive dimension; it has the maximum possible length.
[/example]
This example also warns against a common confusion: Noetherianity stops ascending chains of all ideals, but dimension is measured by finite chains of prime ideals of maximal possible length.
## Failure Modes and Boundary Cases
### Infinite Generation
The Noetherian condition is powerful because it excludes specific pathologies. Seeing what fails outside the Noetherian world helps locate the exact role of the hypothesis in theorems.
The first boundary case is the infinite-variable polynomial ring already encountered. It fails because new independent variables keep producing new ideals, so Hilbert's theorem cannot be extended from finitely many variables to countably many variables without changing the conclusion.
[quotetheorem:10004]
This theorem is a useful diagnostic. Hilbert's basis theorem requires adjoining finitely many variables at a time; infinitely many variables exceed the theorem's scope.
Another boundary case comes from functions. Rings of arbitrary functions often contain ideals defined by infinitely many independent vanishing conditions, so they provide a good testing ground for separating Noetherian and non-Noetherian behaviour.
[example: A Ring of Functions with Many Independent Conditions]
Let $X$ be an infinite set and let $R$ be the ring of all functions $f:X\to k$ to a field $k$, with pointwise addition and multiplication. Choose distinct points $x_1,x_2,\ldots\in X$, and for each $n\ge 1$ define
\begin{align*}
J_n=\{f\in R: f(x)=0 \text{ for all } x\in X\setminus\{x_1,\ldots,x_n\}\}.
\end{align*}
Thus $J_n$ consists exactly of the functions whose support is contained in the finite set $\{x_1,\ldots,x_n\}$.
We first check that $J_n$ is an ideal of $R$. The zero function lies in $J_n$ because it vanishes at every point of $X$. If $f,g\in J_n$ and $x\in X\setminus\{x_1,\ldots,x_n\}$, then $f(x)=0$ and $g(x)=0$, so pointwise addition gives
\begin{align*}
(f+g)(x)=f(x)+g(x)=0+0=0.
\end{align*}
Hence $f+g\in J_n$. If $f\in J_n$ and $h\in R$, then for every $x\in X\setminus\{x_1,\ldots,x_n\}$,
\begin{align*}
(hf)(x)=h(x)f(x)=h(x)0=0.
\end{align*}
Thus $hf\in J_n$, so $J_n$ is an ideal.
The ideals form an ascending chain. If $f\in J_n$ and $x\in X\setminus\{x_1,\ldots,x_n,x_{n+1}\}$, then certainly $x\in X\setminus\{x_1,\ldots,x_n\}$, so $f(x)=0$. Therefore
\begin{align*}
J_n\subseteq J_{n+1}.
\end{align*}
The inclusion is strict. Define $\delta_{n+1}:X\to k$ by
\begin{align*}
\delta_{n+1}(x)=1 \text{ if } x=x_{n+1}, \quad \delta_{n+1}(x)=0 \text{ if } x\ne x_{n+1}.
\end{align*}
If $x\in X\setminus\{x_1,\ldots,x_{n+1}\}$, then $x\ne x_{n+1}$, so $\delta_{n+1}(x)=0$; hence $\delta_{n+1}\in J_{n+1}$. But $x_{n+1}\in X\setminus\{x_1,\ldots,x_n\}$ and
\begin{align*}
\delta_{n+1}(x_{n+1})=1\ne 0,
\end{align*}
so $\delta_{n+1}\notin J_n$. Therefore
\begin{align*}
J_n\subsetneq J_{n+1}
\end{align*}
for every $n\ge 1$.
Thus
\begin{align*}
J_1\subsetneq J_2\subsetneq J_3\subsetneq\cdots
\end{align*}
is an ascending chain of ideals that never stabilises. Hence $R$ fails the ascending chain condition on ideals, so $R$ is not Noetherian.
[/example]
### Artinian Rings
Descending chains are governed by a different finiteness condition. The distinction matters because a ring may stop all ascending chains while still allowing infinite descent, as $\mathbb{Z}$ does through the ideals $(2^n)$. To name the descending analogue, one uses the Artinian condition.
[definition: Artinian Ring]
A ring $R$ is an Artinian ring if every descending chain of ideals
\begin{align*}
I_1 \supset I_2 \supset I_3 \supset \cdots
\end{align*}
stabilises.
[/definition]
Descending-chain control and ascending-chain control look independent at first: one prevents ideals from shrinking forever, while the other prevents them from growing forever. The surprising obstruction is that a ring with all descending chains stopped has so little ideal-theoretic room that infinite ascending chains are also impossible.
[quotetheorem:10005]
This theorem places Artinian rings inside the Noetherian world. The converse is false: $\mathbb{Z}$ is Noetherian, but the descending chain
\begin{align*}
(2) \supsetneq (4) \supsetneq (8) \supsetneq \cdots
\end{align*}
does not stabilise.
[example: Noetherian Does Not Mean Finite]
The ring $\mathbb{Z}$ is Noetherian because every ideal of $\mathbb{Z}$ is principal, and hence every ideal is generated by one element; by *Ideal Characterisation of Noetherian Rings*, this implies that $\mathbb{Z}$ is Noetherian.
This does not mean that $\mathbb{Z}$ is finite as a set: the elements
\begin{align*}
0,1,2,3,\ldots
\end{align*}
are all distinct. It also does not mean that $\mathbb{Z}$ is a field. For example, if $2$ had a multiplicative inverse in $\mathbb{Z}$, there would be an integer $a$ with
\begin{align*}
2a=1.
\end{align*}
But the left-hand side is even and the right-hand side is odd, so no such $a$ exists.
Finally, $\mathbb{Z}$ is not Artinian. For every $n\ge 1$,
\begin{align*}
(2^{n+1})\subseteq (2^n)
\end{align*}
because every element of $(2^{n+1})$ has the form $2^{n+1}a=2^n(2a)$ for some $a\in\mathbb{Z}$. The inclusion is strict: $2^n\in(2^n)$, but if $2^n\in(2^{n+1})$, then $2^n=2^{n+1}a$ for some $a\in\mathbb{Z}$, so cancellation in $\mathbb{Z}$ gives $1=2a$, impossible. Hence
\begin{align*}
(2)\supsetneq(4)\supsetneq(8)\supsetneq\cdots
\end{align*}
is a descending chain of ideals that never stabilises.
Thus Noetherianity is an ascending finiteness condition on ideals, not a statement that the ring is finite, a field, or Artinian.
[/example]
This boundary example is often the quickest way to correct intuition. Noetherian rings are finite in their ideal generation, not necessarily finite in their elements or descending behaviour.
## Beyond and Connected Topics
Noetherian rings are the default setting for much of commutative algebra. The next step is often to study Noetherian modules, exact sequences, localisation, and primary decomposition in a dedicated commutative algebra course such as [Cambridge III Commutative Algebra](/page/Cambridge%20III%20Commutative%20Algebra).
In algebraic geometry, Noetherian rings enter through affine coordinate rings. Hilbert's basis theorem implies that $k[x_1, \ldots, x_n]/I$ is Noetherian, so affine varieties and affine schemes of finite type have finite algebraic descriptions. This connects naturally with [Cambridge II Algebraic Geometry](/page/Cambridge%20II%20Algebraic%20Geometry).
At the undergraduate level, the concept rests on the language of ideals, quotient rings, principal ideal domains, and modules. Those foundations are developed in [Cambridge IB Groups, Rings and Modules](/page/Cambridge%20IB%20Groups%2C%20Rings%20and%20Modules), while the module viewpoint also draws on finite generation and linear algebraic intuition from [Cambridge IB Linear Algebra](/page/Cambridge%20IB%20Linear%20Algebra).
A further direction is dimension theory. Once prime ideals are organised into chains, one can define Krull dimension, local dimension, regular local rings, and depth. These invariants measure the geometry hidden in a commutative Noetherian ring.
Another direction is homological algebra. Over Noetherian rings, finitely generated modules form a robust category, so constructions such as $\operatorname{Tor}_n^R(M,N)$ and $\operatorname{Ext}_R^n(M,N)$ preserve finiteness under useful hypotheses and become tools for measuring singularity and intersection.
## References
Androma, [Cambridge III Commutative Algebra](/page/Cambridge%20III%20Commutative%20Algebra).
Androma, [Cambridge II Algebraic Geometry](/page/Cambridge%20II%20Algebraic%20Geometry).
Androma, [Cambridge IB Groups, Rings and Modules](/page/Cambridge%20IB%20Groups%2C%20Rings%20and%20Modules).
Androma, [Cambridge IB Linear Algebra](/page/Cambridge%20IB%20Linear%20Algebra).
Atiyah and Macdonald, *Introduction to Commutative Algebra* (1969).
Eisenbud, *Commutative Algebra with a View Toward Algebraic Geometry* (1995).
Matsumura, *Commutative Ring Theory* (1986).
Noetherian Ring
Also known as: Noetherian rings, Noetherian commutative ring, ascending chain condition ring, ACC ring, finitely generated ideal ring