Chapter 12
Hard

Ideals & Quotient Rings

00 · Symbol Glossary

$I \trianglelefteq R$I is an ideal of R

Borrowing the normal-subgroup symbol deliberately: ideals play exactly the role in ring theory that normal subgroups played in group theory — the substructures that make quotients possible.

$(a)$Principal ideal generated by a

(a)=aR={ar:rR}(a) = aR = \{ar : r\in R\} — the smallest ideal containing aa, built by absorbing every possible multiple of aa.

$I+J$Sum of two ideals

I+J={i+j:iI,jJ}I+J = \{i+j : i\in I, j\in J\} — the smallest ideal containing both II and JJ.

$R/I$Quotient ring

The ring of cosets a+Ia+I, with (a+I)+(b+I):=(a+b)+I(a+I)+(b+I) := (a+b)+I and (a+I)(b+I):=ab+I(a+I)(b+I) := ab+I. Well-defined precisely when II is an ideal.


01 · Ideals

A subring only needs to be closed under RR's operations. An ideal demands something much stronger from multiplication: it must absorb multiplication by anything in the ambient ring, not merely multiply things that are already inside it.

Definition — Ideal

A subset IRI\subseteq R (commutative ring with unity) is an ideal if: (I1) (I,+)(I,+) is a subgroup of (R,+)(R,+); (I2) for every rRr\in R and aIa\in I, raIra\in I (absorption).

Absorption is strictly stronger than closure

A subring only requires abSab\in S when both a,bSa,b\in S. An ideal requires raIra\in I even when rr comes from all of RR, not just from II. Every ideal is automatically a subring's additive part, but ideals are rarely subrings with unity — see the FailBlock below.


02 · The Ideal Test

Theorem 12.1 — Ideal Test

A nonempty subset IRI\subseteq R is an ideal if and only if: (a) a,bI    abIa,b\in I \implies a-b\in I; (b) rR,aI    raIr\in R,\,a\in I \implies ra\in I.

Proof

(\Rightarrow) I1 gives subtraction-closure (identical to the subgroup-test argument, Theorem 3.1) and I2 is exactly (b).

(\Leftarrow) Condition (a) is precisely the subgroup test (Theorem 3.1) for (I,+)(R,+)(I,+)\le(R,+), giving I1. Condition (b) is I2 directly.


03 · Examples of Ideals

Example — nZ is an ideal of Z

nZ={nk:kZ}n\mathbb{Z} = \{nk : k\in\mathbb{Z}\}: subtraction-closed (already a subgroup, Chapter 03). Absorption: for rZr\in\mathbb{Z} and nknZnk\in n\mathbb{Z}, r(nk)=n(rk)nZr(nk) = n(rk) \in n\mathbb{Z}. nZZn\mathbb{Z}\trianglelefteq\mathbb{Z}.

Example — The trivial ideals

{0}\{0\}: trivially satisfies both conditions. RR itself: trivially satisfies both. Every ring has at least these two ideals, exactly mirroring {e}\{e\} and GG always being normal subgroups.

Definition — Principal Ideal

For aRa\in R, (a):=aR={ar:rR}(a) := aR = \{ar : r\in R\} is the principal ideal generated by aa — verified to be an ideal by the same style of check as nZn\mathbb{Z} above (with aa playing the role of nn).

Common mistake — A proper ideal is never a subring with unity

Is 2Z2\mathbb{Z} (an ideal of Z\mathbb{Z}, since 2Z=(2)2\mathbb{Z}=(2)) a subring with unity? It would need some u2Zu\in2\mathbb{Z} with ua=aua=a for every a2Za\in2\mathbb{Z} — forcing u=1u=1, but 12Z1\notin2\mathbb{Z}. More generally: if a proper ideal IRI\subsetneq R contained 11, absorption (I2) would give r1=rIr\cdot1=r\in I for every rRr\in R, forcing I=RI=R — contradicting properness. A proper ideal never contains 11, and hence is never a subring with unity, even when it happens to be a perfectly good subring otherwise.

Theorem 12.2 — A Field's Only Ideals Are Trivial

If FF is a field (Chapter 17 defines this formally; for now, a commutative ring with unity in which every nonzero element is a unit), then its only ideals are {0}\{0\} and FF.

Proof

Let IFI\trianglelefteq F with I{0}I\neq\{0\}. Pick nonzero aIa\in I. Since FF is a field, aa has an inverse a1Fa^{-1}\in F. By absorption (I2), a1aIa^{-1}\cdot a \in I, i.e. 1I1\in I. Then for any rFr\in F, absorption gives r1=rIr\cdot1=r\in I. So I=FI=F.

A preview of what fields will mean structurally

Theorem 12.2 already hints at why fields will turn out to be exactly the commutative rings whose only quotients are trivial (R/{0}RR/\{0\}\cong R or R/R{0}R/R\cong\{0\}) — there's no "in-between" ideal to quotient by, unlike Z\mathbb{Z}'s rich supply of ideals nZn\mathbb{Z}. This becomes precise in Section 06.


04 · Sums and Intersections of Ideals

Theorem 12.3 — Sums and Intersections of Ideals Are Ideals

If I,JRI,J\trianglelefteq R, then I+J={i+j:iI,jJ}I+J = \{i+j:i\in I,j\in J\} and IJI\cap J are both ideals of RR.

Proof

I+JI+J: nonempty (0+0=0I+J0+0=0\in I+J). Subtraction: (i1+j1)(i2+j2)=(i1i2)+(j1j2)I+J(i_1+j_1)-(i_2+j_2) = (i_1-i_2)+(j_1-j_2) \in I+J, using subtraction-closure inside II and JJ separately. Absorption: r(i+j)=ri+rjI+Jr(i+j) = ri+rj \in I+J, using absorption inside II and inside JJ separately.

IJI\cap J: nonempty (0IJ0\in I\cap J). Subtraction: if a,bIJa,b\in I\cap J, then abIa-b\in I (subtraction-closure in II) and abJa-b\in J (subtraction-closure in JJ), so abIJa-b\in I\cap J. Absorption: if aIJa\in I\cap J and rRr\in R, then raIra\in I and raJra\in J, so raIJra\in I\cap J.

Example — Sums and intersections in Z

4Z+6Z4\mathbb{Z}+6\mathbb{Z}: contains 44 and 66, hence (being an ideal, closed under subtraction) contains 64=26-4=2, hence contains (2)=2Z(2)=2\mathbb{Z}; conversely every element of 4Z+6Z4\mathbb{Z}+6\mathbb{Z} is a multiple of gcd(4,6)=2\gcd(4,6)=2. So 4Z+6Z=2Z4\mathbb{Z}+6\mathbb{Z}=2\mathbb{Z}sum of ideals corresponds to gcd.

4Z6Z4\mathbb{Z}\cap6\mathbb{Z}: multiples of both 44 and 66 are exactly multiples of lcm(4,6)=12\mathrm{lcm}(4,6)=12. So 4Z6Z=12Z4\mathbb{Z}\cap6\mathbb{Z}=12\mathbb{Z}intersection corresponds to lcm, a connection that resurfaces directly in Chapter 16's Chinese Remainder Theorem.


05 · Quotient Rings

Theorem 12.4 — R/I Is a Ring When I Is an Ideal

Let IRI\trianglelefteq R. Then (a+I)+(b+I):=(a+b)+I(a+I) + (b+I) := (a+b)+I and (a+I)(b+I):=ab+I(a+I)(b+I) := ab+I are well-defined operations on the set of cosets R/IR/I, and R/IR/I is a commutative ring with unity 1+I1+I.

Proof

Since (R,+)(R,+) is abelian, (I,+)(I,+) is automatically normal (Chapter 05, every subgroup of an abelian group is normal), so addition on R/IR/I is well-defined and (R/I,+)(R/I,+) is an abelian group by Theorem 5.3 — this part is pure group theory, already proven.

Multiplication well-defined. Suppose a+I=a+Ia+I=a'+I and b+I=b+Ib+I=b'+I, so a=a+i1a'=a+i_1, b=b+i2b'=b+i_2 for i1,i2Ii_1,i_2\in I (Theorem 4.1, additive version). Then:

ab=(a+i1)(b+i2)=ab+ai2+i1b+i1i2a'b' = (a+i_1)(b+i_2) = ab + ai_2 + i_1b + i_1i_2

Each of ai2ai_2, i1bi_1b, i1i2i_1i_2 lies in II: ai2Iai_2\in I and i1bIi_1b\in I by absorption (I2), and i1i2Ii_1i_2\in I by absorption applied to i1Ii_1\in I multiplied by the ring element i2i_2 (or simply because II is closed under multiplication by anything in RR, in particular by elements of IRI\subseteq R). So abab=ai2+i1b+i1i2Ia'b' - ab = ai_2+i_1b+i_1i_2 \in I (sum of elements of II, using subtraction-closure repeatedly), giving ab+I=ab+Ia'b'+I = ab+I.

Ring axioms. Associativity, distributivity, and commutativity of the coset multiplication are inherited directly from the corresponding properties in RR (the same style of verification as Theorem 5.3's group axioms for G/NG/N). 1+I1+I is the multiplicative identity: (1+I)(a+I)=a+I(1+I)(a+I) = a+I.

Example — $\mathbb{Z}/n\mathbb{Z}$ as a ring, confirmed

The quotient ring construction applied to I=nZZI=n\mathbb{Z}\trianglelefteq\mathbb{Z} gives exactly Zn\mathbb{Z}_n with its ring structure from Chapter 11 — Theorem 12.4 confirms that the well-definedness proof carried out by hand in Theorem 1.5 was a special case of this general machinery.


06 · Prime and Maximal Ideals

Definition — Prime and Maximal Ideals

A proper ideal IRI\subsetneq R is prime if abI    aIab\in I \implies a\in I or bIb\in I. It is maximal if there is no ideal JJ with IJRI\subsetneq J\subsetneq R (no proper ideal strictly between II and RR).

Theorem 12.5 — R/I Is a Domain iff I Is Prime

R/IR/I is an integral domain if and only if II is a prime ideal.

Proof

R/IR/I has 1+I0+I1+I\neq0+I exactly when 1I1\notin I, i.e. IRI\neq R: proper. R/IR/I has no zero divisors exactly when: (a+I)(b+I)=0+I    a+I=0+I(a+I)(b+I) = 0+I \implies a+I=0+I or b+I=0+Ib+I=0+I, i.e. abI    aIab\in I \implies a\in I or bIb\in I. This is precisely the definition of II being prime. So "R/IR/I is a domain" and "II is prime" are, unpacked, the identical statement.

Theorem 12.6 — R/I Is a Field iff I Is Maximal

R/IR/I is a field if and only if II is a maximal ideal.

Proof

(\Leftarrow) Suppose II is maximal. R/IR/I is a nonzero commutative ring with unity (since II is proper). Take any nonzero a+Ia+I, i.e. aIa\notin I. Consider J=I+(a)J=I+(a) (an ideal, by Theorem 12.3, since (a)(a) is an ideal). Since a=a1(a)Ja=a\cdot1\in(a)\subseteq J but aIa\notin I, JIJ\supsetneq I (strictly larger). By maximality of II, the only ideal strictly containing II is RR itself, so J=RJ=R. In particular 1J=I+(a)1\in J=I+(a), so 1=i+ar1=i+ar for some iIi\in I, rRr\in R. Then ar1=iIar-1 = -i\in I, so (a+I)(r+I)=ar+I=1+I(a+I)(r+I) = ar+I = 1+I: a+Ia+I has inverse r+Ir+I. Since a+Ia+I was an arbitrary nonzero element, R/IR/I is a field.

(\Rightarrow) Suppose R/IR/I is a field; in particular IRI\neq R (proper, since 1+I0+I1+I\neq0+I). Let JJ be an ideal with IJRI\subsetneq J\subseteq R; pick aJIa\in J\setminus I. Since aIa\notin I, a+I0+Ia+I\neq0+I in R/IR/I; as R/IR/I is a field, a+Ia+I has an inverse r+Ir+I: (a+I)(r+I)=1+I(a+I)(r+I)=1+I, i.e. ar1IJar-1\in I \subseteq J. Since aJa\in J and JJ absorbs multiplication by rRr\in R, arJar\in J. Then 1=ar(ar1)1 = ar - (ar-1) is a difference of two elements of JJ (arJar\in J and ar1Jar-1\in J), so 1J1\in J (subtraction-closure). By absorption, Jr1=rJ\ni r\cdot1=r for every rRr\in R, so J=RJ=R. No ideal sits strictly between II and RR: II is maximal.

Example — Maximal ideals of Z are exactly (p) for prime p

Z/(n)=Zn\mathbb{Z}/(n) = \mathbb{Z}_n is a field exactly when every nonzero element is a unit — by Theorem 11.4, that happens iff gcd(a,n)=1\gcd(a,n)=1 for every a{1,,n1}a\in\{1,\ldots,n-1\}, which holds iff nn is prime. By Theorem 12.6, (n)(n) is maximal in Z\mathbb{Z} exactly when nn is prime — matching the elementary number-theoretic fact that primes have no nontrivial factorizations, now derived as a ring-theoretic statement.

Maximal ideals are always prime

Since fields are always integral domains (a field has no zero divisors: if ab=0ab=0 with a0a\neq0, multiply by a1a^{-1} to get b=0b=0), Theorem 12.6 combined with Theorem 12.5 shows every maximal ideal is automatically prime — R/IR/I a field     \implies R/IR/I a domain     \implies II prime. The converse can fail: (0)(0) is prime in Z\mathbb{Z} (since Z\mathbb{Z} has no zero divisors) but not maximal (since Z\mathbb{Z}\subsetneq nothing is needed — rather, (0)(2)Z(0)\subsetneq(2)\subsetneq\mathbb{Z} shows (0)(0) is not maximal).


07 · Exercises

EXERCISE 12.1

Apply Theorem 12.1 directly: check subtraction-closure and absorption for multiples of 5.

I=5ZI=5\mathbb{Z}. Subtraction: 5j5k=5(jk)5Z5j-5k=5(j-k)\in5\mathbb{Z}. Absorption: for rZr\in\mathbb{Z}, r(5k)=5(rk)5Zr(5k)=5(rk)\in5\mathbb{Z}. Both conditions hold: 5ZZ5\mathbb{Z}\trianglelefteq\mathbb{Z}.

Verify using Theorem 12.1 that 5Z5\mathbb{Z} is an ideal of Z\mathbb{Z}.

EXERCISE 12.2

Apply the gcd/lcm identification from Section 04's example.

6Z+10Z=gcd(6,10)Z=2Z6\mathbb{Z}+10\mathbb{Z} = \gcd(6,10)\mathbb{Z} = 2\mathbb{Z}. 6Z10Z=lcm(6,10)Z=30Z6\mathbb{Z}\cap10\mathbb{Z} = \mathrm{lcm}(6,10)\mathbb{Z} = 30\mathbb{Z}.

Find 6Z+10Z6\mathbb{Z}+10\mathbb{Z} and 6Z10Z6\mathbb{Z}\cap10\mathbb{Z}.

EXERCISE 12.3

Apply Theorem 12.5: is (4)(4) a prime ideal of Z\mathbb{Z}? Find a specific product that breaks primality if not.

2×2=4(4)2\times2=4\in(4), but 2(4)2\notin(4). So (4)(4) is not prime: the implication "abI    aIab\in I\implies a\in I or bIb\in I" fails with a=b=2a=b=2.

By Theorem 12.5, since (4)(4) is not prime, Z/(4)=Z4\mathbb{Z}/(4)=\mathbb{Z}_4 is not an integral domain — confirmed directly, since [2][2]=[0][2][2]=[0] in Z4\mathbb{Z}_4 with [2][0][2]\neq[0] (Chapter 11, Exercise 11.3's sibling fact).

Is (4)(4) a prime ideal of Z\mathbb{Z}? Use Theorem 12.5 to conclude whether Z4\mathbb{Z}_4 is an integral domain.

EXERCISE 12.4

Apply Theorem 12.6 directly, then recall which Zn\mathbb{Z}_n are fields.

77 is prime, so by Section 06's Example, (7)(7) is maximal in Z\mathbb{Z}. By Theorem 12.6, Z/(7)=Z7\mathbb{Z}/(7) = \mathbb{Z}_7 is a field — every nonzero element has an inverse (confirmed directly for Z5\mathbb{Z}_5 in Exercise 1.5, and the same argument works for any prime).

Using Theorem 12.6, determine whether Z7\mathbb{Z}_7 is a field.


08 · Chapter Summary

ConceptStatement
IdealAdditive subgroup that absorbs multiplication by all of RR (I1, I2)
Ideal testSubtraction-closed + absorbs multiplication (Thm 12.1)
Principal ideal(a)=aR(a)=aR
Field's idealsOnly {0}\{0\} and FF (Thm 12.2)
Sum, intersection of idealsBoth are ideals (Thm 12.3); correspond to gcd, lcm in Z\mathbb{Z}
Quotient ring R/IR/IWell-defined ring structure iff II is an ideal (Thm 12.4)
Prime idealabI    aIab\in I\implies a\in I or bIb\in I
Maximal idealNo proper ideal strictly between II and RR
R/IR/I domain     I\iff I primeThm 12.5
R/IR/I field     I\iff I maximalThm 12.6

Next: Chapter 13 — Ring Homomorphisms builds the ring-theoretic analog of Chapter 06's homomorphism and isomorphism theorems, with ideals playing the role kernels of group homomorphisms played for normal subgroups.