Rings — Axioms, Subrings, Ideals

A ring is a set equipped with two operations — addition and multiplication — where addition forms an abelian group, multiplication is associative, and the two are linked by distributivity. Subrings are subsets closed under both operations; ideals are special subrings that absorb multiplication from the whole ring and are precisely the kernels of ring homomorphisms, enabling quotient constructions. Used everywhere in algebra, number theory, algebraic geometry, and coding theory. Formalized by Dedekind, Hilbert, Noether, and Fraenkel (1870s–1920s).

A ring $(R, +, \cdot)$ is an abelian group under $+$ together with an associative multiplication that distributes over addition. Rings generalize the familiar arithmetic of the integers $Z$ , abstracting the essential interplay of two operations while permitting wild new behaviors like zero divisors and noncommutativity. They matter because they capture the algebraic structure of polynomials, matrices, modular arithmetic, functions, and number fields — virtually all of computational and abstract algebra lives inside ring theory. The surprising fact: ideals, not subrings, are the "normal subgroups" of ring theory — only they yield well-defined quotients — and a single ring can have rich ideal structure invisible from its element-by-element description (e.g., the polynomial ring $k [x]$ has exactly one maximal chain of ideals per irreducible polynomial).

The Core Insight

A ring fuses two algebraic structures on one set: an additive abelian group and a multiplicative monoid (or semigroup), glued by distributivity. This single bridge — that multiplication respects addition — is what makes the whole theory cohere. The consequence is that the right notion of substructure for quotients is not the subring but the ideal: a subgroup that is also a "multiplicative black hole," absorbing products with arbitrary ring elements. Ideals are exactly the kernels of homomorphisms, and the First Isomorphism Theorem makes them the engine of all quotient constructions.

Formal Statement

Ring

A ring is a triple $(R, +, \cdot)$ where $R$ is a set with two binary operations satisfying:

Additive axioms —

(R, +)

is an abelian group:

(a + b) + c = a + (b + c) for all a, b, c \in R

⟳associativity of addition; this is an axiom, not a theorem.

a + b = b + a, a + 0 = a, a + (- a) = 0

⟳commutativity, additive identity

0

, and additive inverses — the abelian group structure.

Multiplicative axiom —

(R, \cdot)

is a semigroup:

(a \cdot b) \cdot c = a \cdot (b \cdot c) for all a, b, c \in R

⟳associativity of multiplication; postulated.

Bridge axioms — distributivity links the two:

a \cdot (b + c) = a \cdot b + a \cdot c, (a + b) \cdot c = a \cdot c + b \cdot c

⟳left and right distributive laws; both required since multiplication need not commute.

Here $0$ is the additive identity, $- a$ the additive inverse of $a$ . A ring is commutative if $a \cdot b = b \cdot a$ for all $a, b$ , and is a ring with identity (or unital) if there exists $1 \in R$ with $1 \cdot a = a \cdot 1 = a$ for all $a$ .

Conventions differ. Many authors (Bourbaki, most modern texts) require a multiplicative identity $1$ in the definition of "ring," reserving rng (ring without identity, the "missing i") for the non-unital case. Others (older texts, some ring-theory specialists) do not. Throughout this article, "ring" allows the unital case as standard, and we note where $1$ is needed. Also: $1 = 0$ forces $R = {0}$ , the zero ring.

Elementary Consequences of the Axioms

In any ring $R$ , for all $a, b \in R$ :

a \cdot 0 = 0 \cdot a = 0

⟳

a \cdot 0 = a \cdot (0 + 0) = a \cdot 0 + a \cdot 0

; cancel in the additive group.

a \cdot (- b) = (- a) \cdot b = - (a \cdot b)

⟳

a \cdot b + a \cdot (- b) = a \cdot (b + (- b)) = a \cdot 0 = 0

, so

a \cdot (- b)

is the additive inverse of

a \cdot b

(- a) \cdot (- b) = a \cdot b

⟳apply [6] twice —

(- a) (- b) = - (a (- b)) = - (- (ab)) = ab

Subring

A subset $S \subseteq R$ is a subring if $S$ is itself a ring under the restricted operations of $R$ . Concretely, $S$ is a subring iff it is nonempty and closed under subtraction and multiplication:

a, b \in S ⟹ a - b \in S and a \cdot b \in S

⟳closure under subtraction makes

(S, +)

a subgroup (subgroup criterion); multiplicative closure inherits associativity and distributivity automatically.

If $R$ is unital, a unital subring additionally requires $1_{R} \in S$ . (Without this, $S$ may have a different identity, or none — see the example below.)

Ideal

A subset $I \subseteq R$ is a two-sided ideal if it is an additive subgroup that absorbs multiplication on both sides:

a, b \in I ⟹ a - b \in I; r \in R, a \in I ⟹ r \cdot a \in I and a \cdot r \in I

⟳the absorption condition

R I \subseteq I

and

I R \subseteq I

is strictly stronger than mere closure under products of elements of

I

A left ideal requires only $r \cdot a \in I$ ; a right ideal only $a \cdot r \in I$ . In a commutative ring all three notions coincide, and we simply say ideal. Every ideal is a subring (often non-unital), but most subrings are not ideals.

Think of an ideal as a set of "negligible" elements that stays negligible no matter what you multiply it by. In $Z$ , the multiples of $5$ form the ideal $5 Z$ : multiply any multiple of $5$ by anything and you still get a multiple of $5$ . That absorption is exactly what lets you "set $5 = 0$ " consistently and build $Z /5 Z$ . A subring like $Z \subset Q$ has no such absorbing property — multiplying the integer $1$ by $\frac{1}{2}$ escapes $Z$ — which is why $Z$ is not an ideal of $Q$ and you cannot form " $Q / Z$ " as a ring.

Why It Holds / Derivation

Ideals are exactly the kernels of ring homomorphisms

Let $φ : R \to S$ be a ring homomorphism (so $φ (a + b) = φ (a) + φ (b)$ , $φ (ab) = φ (a) φ (b)$ ). Define $ker φ = {a \in R : φ (a) = 0}$ .

Step 1 — Kernel is an additive subgroup. If

a, b \in ker φ

then

φ (a - b) = φ (a) - φ (b) = 0 - 0 = 0

, so

a - b \in ker φ

. Thus

ker φ

is a subgroup of

(R, +)

. Step 2 — Kernel absorbs multiplication. For

r \in R

a \in ker φ

φ (r \cdot a) = φ (r) φ (a) = φ (r) \cdot 0 = 0

⟳homomorphism property plus [5]; symmetrically

φ (a \cdot r) = 0

Hence $r \cdot a, a \cdot r \in ker φ$ , so $ker φ$ is a two-sided ideal.

Step 3 — Every ideal is a kernel. Conversely, given a two-sided ideal

I \subseteq R

, form the quotient group

R / I

with cosets

a + I

. Define multiplication on cosets by

(a + I) (b + I) := ab + I

⟳well-definedness needs the absorption axiom — if

a^{'} = a + i

b^{'} = b + j

with

i, j \in I

, then

a^{'} b^{'} = ab + (ib + aj + ij)

, and each of

ib, aj, ij \in I

by [9].

This multiplication is associative and distributive (inherited from $R$ ), so $R / I$ is a ring. The quotient map $π : R \to R / I$ , $π (a) = a + I$ , is a surjective homomorphism with $ker π = I$ . Therefore every ideal arises as a kernel. $■$

First Isomorphism Theorem for Rings

If $φ : R \to S$ is a ring homomorphism, then $ker φ$ is an ideal of $R$ , $im φ$ is a subring of $S$ , and

R / ker φ ≅ im φ

⟳the map

\overset{φ}{ˉ} (a + ker φ) = φ (a)

is well-defined (constant on cosets), bijective onto the image, and respects both operations.

The hierarchy of subsets: subgroup ⊃ subring ⊃ ideal, with the absorption condition shown

Emmy Noether | 1882–1935

Though Dedekind introduced ideals (in number rings) and Hilbert proved his Basis Theorem, it was Emmy Noether who, in her 1921 paper Idealtheorie in Ringbereichen, founded abstract ring theory as we know it. She recast ideal theory in fully axiomatic terms, isolated the ascending chain condition (Noetherian rings), and proved that primary decomposition holds in any commutative ring satisfying it. Her structural, axiomatic viewpoint — "it is all in Noether" as her students said — replaced explicit computation with conceptual reasoning and shaped 20th-century algebra. The First Isomorphism Theorem and the centrality of ideals over subrings are her intellectual legacy.

Building Intuition

Three mental models clarify the axioms:

$Z$ is the prototype. Everything about general rings is a calibrated departure from the integers: commutative? (matrices say no), has $1$ ? (even integers $2 Z$ say no), no zero divisors? ( $Z /6 Z$ says no, since $2 \cdot 3 = 0$ ).

Ideals = "things we declare zero." The quotient $R / I$ is the ring you get by forcing every element of $I$ to vanish. Absorption guarantees this declaration is consistent with multiplication.

Subring vs ideal = "self-contained vs sticky." A subring is a smaller ring sitting inside; an ideal is sticky — it grabs products with anything in the ambient ring. The integers inside the rationals are self-contained but not sticky, hence a subring but not an ideal.

Worked Examples

Ideals of $\mathbb{Z}$ and a quotient computation

Given: The ring

Z

. Find all its ideals, then compute

Z /12 Z

and identify zero divisors. Step 1 — Classify ideals. Every ideal

I \subseteq Z

is an additive subgroup, and the subgroups of

(Z, +)

are exactly

n Z

for

n \geq 0

. Each

n Z

absorbs multiplication (

m \cdot (nk) = n (mk) \in n Z

), so:

{ideals of Z} = {n Z : n = 0, 1, 2, 3, \dots}

⟳subgroups of

Z

are cyclic; cyclic additive subgroups are automatically ideals because

Z

is commutative with the absorbing multiple structure.

Step 2 — Form the quotient. With

I = 12 Z

Z /12 Z = {\overline{0}, \overline{1}, \overline{2}, \dots, \overline{11}}, ∣ Z /12 Z ∣ = 12

⟳cosets

a + 12 Z

are determined by

a mod 12

Step 3 — Zero divisors.

\overline{a} \cdot \overline{b} = \overline{0}

iff

12 ∣ ab

. E.g.

\overline{3} \cdot \overline{4} = \overline{12} = \overline{0}

. The zero divisors are the classes

\overline{a}

with

g cd (a, 12) > 1

{\overline{2}, \overline{3}, \overline{4}, \overline{6}, \overline{8}, \overline{9}, \overline{10}}

. Boxed answer:

Z /12 Z has 12 elements; zero divisors are the classes with g cd (a, 12) > 1.

A subring that is not an ideal, and an ideal in a polynomial ring

Given: (a) Show

Z \subset Q

is a subring but not an ideal. (b) In

R = R [x]

, describe the ideal

I = (x^{2} - 1)

and the quotient

R / I

. Part (a).

Z

is closed under subtraction and multiplication, so it is a subring of

Q

. But take

r = \frac{1}{2} \in Q

and

a = 1 \in Z

r \cdot a = \frac{1}{2} \cdot 1 = \frac{1}{2} \in / Z

⟳absorption fails — a single product escapes the subset — so

Z

cannot be an ideal of

Q

Part (b). The principal ideal is

I = (x^{2} - 1) = {f (x) (x^{2} - 1) : f \in R [x]}

. In the quotient we impose

x^{2} - 1 = 0

, i.e.

x^{2} = 1

. Every polynomial reduces modulo

x^{2} - 1

to a remainder of degree

< 2

R / I ≅ {a + b x + I : a, b \in R}, x^{2} = 1 in the quotient

⟳division algorithm in

R [x]

gives unique remainder

a + b x

; the quotient is a 2-dimensional

R

-algebra.

Since $x^{2} - 1 = (x - 1) (x + 1)$ with coprime factors, the Chinese Remainder Theorem gives:

R [x] / (x^{2} - 1) ≅ R [x] / (x - 1) \times R [x] / (x + 1) ≅ R \times R

⟳

(x - 1)

and

(x + 1)

are comaximal since

\frac{1}{2} (x + 1) - \frac{1}{2} (x - 1) = 1

; apply CRT for rings.

Boxed answer:

Z is a subring (not ideal) of Q; R [x] / (x^{2} - 1) ≅ R \times R .

Common Misconceptions

Pitfall 1: "Every subring is an ideal." False. Ideals require absorption by the whole ring, a far stronger condition.

Z \subset Q

is the standard counterexample. Pitfall 2: "An ideal contains $1$ ." If a (proper) ideal

I

contains

1

, then

r = r \cdot 1 \in I

for all

r

, forcing

I = R

. So a proper ideal never contains the identity — indeed never contains any unit. Pitfall 3: " $ab = 0 \Rightarrow a = 0$ or $b = 0$ ." This is the cancellation/integral-domain property, NOT a ring axiom. Rings can have zero divisors (see

Z /12 Z

). Pitfall 4: "Multiplication must be commutative / have an identity." Neither is a ring axiom.

M_{n} (R)

is noncommutative;

2 Z

has no identity.

A non-principal ideal

In $Z$ every ideal is principal (generated by one element). But in $R = Z [x]$ the ideal $I = (2, x) = {2 f (x) + xg (x) : f, g \in Z [x]}$ — the polynomials with even constant term — is not principal. If $I = (d)$ for some $d$ , then $d ∣ 2$ and $d ∣ x$ . From $d ∣ 2$ , $d \in {\pm 1, \pm 2}$ ; from $d ∣ x$ , $d$ must be a unit, so $d = \pm 1$ and $I = R$ . But $1 \in / I$ (its constant term is odd). Contradiction. Thus $Z [x]$ is not a principal ideal domain, showing principality is a special, not generic, feature.

A "subring" with a different identity

Inside $R = Z \times Z$ (componentwise operations, identity $(1, 1)$ ), the subset $S = Z \times {0}$ is closed under subtraction and multiplication, hence a subring. Yet its multiplicative identity is $(1, 0) \neq = (1, 1) = 1_{R}$ . So a subring may possess an identity different from the ambient ring's — a reason many texts demand $1_{S} = 1_{R}$ for "unital subring."

Connections

Ring theory sits at a crossroads:

Groups — Axioms and Subgroups: the additive structure of a ring is an abelian group, and ideals play the role for rings that normal subgroups play for groups (both are exactly the kernels enabling quotients).
Fields and Field Extensions: a field is a commutative ring with $1$ where every nonzero element is a unit; equivalently, a ring whose only ideals are $0$ and itself.
Quotient Rings and Isomorphism Theorems: built directly from ideals via the construction in [11]–[12].
Maximal and Prime Ideals: $R / I$ is a field iff $I$ is maximal; $R / I$ is an integral domain iff $I$ is prime — translating ring-theoretic properties into ideal-theoretic ones.
Modules over a Ring: ideals are precisely the submodules of $R$ viewed as a module over itself.

Real-World Applications

Cryptography (RSA). Works in the ring $Z / n Z$ with $n = pq$ ; security rests on the ideal/factorization structure and the units group $(Z / n Z)^{\times}$ .
Error-correcting codes. Cyclic codes are ideals in the quotient ring $F_{q} [x] / (x^{n} - 1)$ ; generator polynomials are ideal generators (Reed–Solomon, BCH codes).
Computer algebra systems. Gröbner bases compute with ideals in $k [x_{1}, \dots, x_{n}]$ to solve polynomial systems (used in Maple, Macaulay2, SageMath).
Algebraic geometry / robotics. The ideal–variety correspondence (Nullstellensatz) translates geometric problems (robot kinematics, motion planning) into ideal membership.
Number theory. Unique factorization fails in rings like $Z [- 5]$ but is restored for ideals — the historical origin of the concept (Kummer's "ideal numbers," Dedekind's ideals).
Signal processing. Polynomial rings and their quotients model convolution and the Number-Theoretic Transform used in fast integer multiplication.
Quantum computing & coding. Stabilizer codes and lattice-based post-quantum cryptography (e.g. Ring-LWE) operate in quotient polynomial rings $Z_{q} [x] / (x^{n} + 1)$ .

Historical Development

1847 — Ernst Kummer introduces "ideal complex numbers" to rescue unique factorization in cyclotomic integers while studying Fermat's Last Theorem.
1871 — Richard Dedekind replaces Kummer's ideal numbers with actual sets (ideals) in his supplements to Dirichlet's lectures, defining ideals in rings of algebraic integers.
1882 — Dedekind and Weber develop ideal theory for algebraic function fields, foreshadowing algebraic geometry.
1888 — David Hilbert proves the Basis Theorem: every ideal in $k [x_{1}, \dots, x_{n}]$ is finitely generated (Noetherian property avant la lettre).
1897 — Hilbert's Zahlbericht systematizes algebraic number theory in ideal-theoretic language.
1914 — Abraham Fraenkel gives an early axiomatic definition of (a class of) rings; the German term Ring (from Hilbert) becomes standard.
1921 — Emmy Noether's Idealtheorie in Ringbereichen founds abstract commutative ring theory; ascending chain condition and primary decomposition.
1927 — Emil Artin and Noether develop the structure theory of rings with chain conditions (Artinian rings, Wedderburn–Artin).
1930 — van der Waerden's Moderne Algebra, distilling Noether's and Artin's lectures, canonizes the modern axiomatic treatment for generations of students.

Subtleties & Limits

Several deep questions surround even the basic theory:

Köthe's Conjecture (1930, open). If a ring has no nonzero nil one-sided ideal, does it have no nonzero nil two-sided ideal? Equivalently, is the sum of two nil left ideals always nil? Unresolved for nearly a century — one of the oldest open problems in noncommutative ring theory.

Invariant Basis Number subtleties. For general (noncommutative) rings $R$ , free modules can satisfy $R^{m} ≅ R^{n}$ with $m \neq = n$ ; characterizing such "non-IBN" rings remains an active study.

Computational hardness. Ideal membership in $k [x_{1}, \dots, x_{n}]$ is decidable (Gröbner bases) but EXPSPACE-complete in general; practical computation is the bottleneck of computational algebraic geometry.

Definitional drift. The "ring vs rng" convention is genuinely unsettled across the literature; results like "every ring embeds in a unital ring" (the Dorroh extension) exist precisely to manage the non-unital case, a reminder that the axioms one chooses have real downstream consequences.