Motivation
Reducing integers modulo preserves both addition and multiplication. A ring quotient generalizes that construction, but its kernel must interact correctly with both operations. Additive subgroups alone are insufficient: multiplying two congruent representatives by an arbitrary ring element must preserve their congruence. This extra condition is exactly the ideal condition.
The chapter uses elementary group quotients and integer Bézout identities. Unless explicitly noted, every ring here is commutative with an identity , and homomorphisms preserve identity. Quotients by proper ideals remain within this convention. A quotient by the whole ring is the zero ring, which we name when needed but do not call a field or an integral domain.
Definitions
Definition 3.1.1
(Rings, units, and zero divisors)
A ring is an abelian group under addition, equipped with associative multiplication distributing over addition on both sides and having identity . It is commutative when . A unit has a multiplicative inverse in . A zero divisor is a nonzero for which with some nonzero . An integral domain is a nonzero commutative ring with identity and no zero divisors. A field is such a ring in which every nonzero element is a unit.
An element can fail to be a unit without being a zero divisor: the integer two in is an example. In a domain, cancellation is valid when , since . In a finite domain, multiplication by a nonzero is injective and therefore surjective, so some satisfies . Thus every finite domain is a field. Finiteness is essential; is an infinite domain and not a field.
Definition 3.1.2
(Ideals and generated ideals)
An ideal is an additive subgroup such that whenever and . It is proper when . The principal ideal is . More generally consists of all sums .
These sets are ideals by the distributive laws. They are the smallest ideals containing the indicated generators: every ideal containing the generators contains all those sums. An ideal equals exactly when it contains , and any ideal containing a unit contains . Hence a field has only the ideals and itself.
The sum and intersection are ideals. The product consists of all finite sums of products with . Allowing finite sums is necessary; single products need not form an ideal. Always .
Theorems
Theorem 3.1.3
(Ring quotients and the first isomorphism theorem)
For an ideal , the operations and make a commutative ring with identity . If is a unital ring homomorphism, its kernel is an ideal and as rings.
Theorem 3.1.4
(Prime and maximal ideals)
A proper ideal is prime when forces or . It is prime exactly when is an integral domain. A proper ideal is maximal when no ideal lies strictly between and . It is maximal exactly when is a field. Consequently every maximal ideal is prime.
Proof sketch or proof idea
Proof
(Well-defined quotient operations and the kernel)
For quotient multiplication, change representatives to and , with . The difference from is , in by absorption. This proves well-definedness; addition uses additive closure. Ring identities descend from . The kernel of a ring homomorphism is an additive subgroup, and whenever is in it. The induced map is a group isomorphism by the earlier theorem, and preserves multiplication and identity, proving the ring version.
Proof
(Prime and maximal ideals through their quotients)
A coset is zero exactly when . Thus absence of nonzero zero divisors in is precisely the prime-ideal implication. Properness ensures the quotient identity is nonzero. This proves the prime characterization in both directions.
If is maximal and , the ideal strictly contains , hence equals . Write with . In the quotient this becomes , giving an inverse for every nonzero coset. Conversely, if is a field and an ideal strictly contains , choose . An inverse coset gives for some , so and . Finally, a field has no zero divisors: multiply by if . The implication maximal implies prime now follows.
Chinese remainders as independent coordinates
Theorem 3.1.5
(Chinese remainder theorem for two ideals)
If , then and
The product ring has coordinatewise operations. The condition is called comaximality.
Proof
(Comaximality gives surjectivity and )
Consider , . Its kernel is without any extra hypothesis. Comaximality supplies with . To realize an arbitrary target , take . Modulo , and , giving ; modulo , and , giving . This proves surjectivity, rather than merely assuming the coordinates are independent.
If , then . The product belongs to because ; likewise because . Therefore , and the reverse inclusion was already established. Apply the ring isomorphism theorem to finish the proof. Repeating this argument yields a finite product for pairwise comaximal ideals; this chapter does not use an infinite-product version.
Proof X-ray
One identity does both jobs
Comaximality is used exactly once, to produce and with ; the rest of the proof is bookkeeping with that identity. Since , the element is modulo and modulo , while is the reverse, so is assembled from two coordinate elements with the prescribed entries. Multiplying by the same identity writes as a sum of products of an element of with an element of . Without comaximality both conclusions can fail: in , while .
For integers , exactly when , by Bézout. The theorem therefore gives under the coprimality hypothesis. Both addition and multiplication are preserved, not merely the count of residue classes.
Which ideals survive a quotient?
For an ideal of , ideals of correspond to ideals of containing . The correspondence sends to and an ideal of the quotient to its preimage under the projection. The additive subgroup correspondence was proved for groups. To check the extra ideal property, if and , their product is . Conversely, if an element maps into , multiplying it by any ring element still maps into , so the preimage is an ideal. The inverse identities follow from surjectivity exactly as in the subgroup case.
This also proves when . The map is ; it is a ring isomorphism by the first theorem. For example, start with integers modulo twelve, and then identify all even classes with zero. The result is arithmetic modulo two, since the ideal generated by is the image of , which contains . It is easy to confuse the size of the intermediate ideal with the size of its quotient; the ideal has six elements, while the final quotient has two.
Products can create zero divisors
In a product of two nonzero commutative rings, the elements and are nonzero and multiply to zero. The product ring is therefore never a domain. This gives a structural explanation of why an integer modulus with two coprime factors produces zero divisors: its Chinese remainder decomposition has two nonzero coordinates. For instance, modulo fifteen, the class of six corresponds to modulo three and five, and the class of ten corresponds to . Their product is zero modulo fifteen.
An element satisfying is called an idempotent. The two coordinate elements are idempotents whose sum is one and whose product is zero. In a domain, the equation forces or . Thus a nontrivial idempotent is another immediate obstruction to being a domain. These computations explain why a product decomposition changes multiplication so substantially, despite preserving a familiar additive size.
Concept lensStructural
Product decompositions as idempotents
A commutative ring splits as a product of two nonzero rings exactly when it has an idempotent . One direction is the coordinate element . Conversely, through , where is a ring with identity ; the cross terms vanish because . The Chinese remainder theorem is an instance: with as in its proof, , so is an idempotent of , and it corresponds to .
A nilpotent element satisfies for some positive integer . Every prime ideal contains every nilpotent element: repeatedly apply the prime implication to until obtaining . In a domain no nonzero element is nilpotent. The ring has the nonzero nilpotent , whereas has zero divisors but no nonzero nilpotent. For the latter assertion, if six divides , both primes two and three divide , so six divides . Thus having zero divisors is a weaker condition than having a nonzero nilpotent. The distinction will matter when studying polynomial quotients with repeated factors.
The integers inside a ring
What does an integer coefficient mean when the elements of a ring are not themselves integers? For , write for the sum of copies of , put , and set . The distributive laws give
For nonnegative integers, the second equality expands a product into copies of one. Additive inverses extend it to all signs. Consequently , , is a unital ring homomorphism. It is the only such homomorphism: preserving one forces the images of positive integers by addition, then zero and negative integers by additive inverses. Integer coefficients therefore have a canonical interpretation even when that interpretation is not injective.
The kernel records exactly which integer coefficients become zero. If it has a positive element, let be its least positive element. Dividing any as , with , shows , so . Conversely every multiple of is killed. Thus . If there is no positive kernel element, the kernel is , since any negative element would have a positive additive inverse. Set in this case.
This is the characteristic, written . Since , it is either zero or at least two. The first isomorphism theorem identifies the image of with for , and with for . The image is the prime subring: it is contained in every subring having the same identity as , because such a subring must contain every integer multiple of one. The adjective does not assert that this subring is a prime ideal or an integral domain.
If is a domain and with , then and are nonzero by minimality, but their product is . Hence a domain has characteristic zero or a prime. The converse fails: has characteristic two, yet its nonzero elements and multiply to zero. Characteristic describes integer coefficients, rather than all possible zero divisors. Characteristic zero likewise supplies a copy of , not automatically a copy of ; the ring itself does not contain an inverse of two.
Theorem 3.1.6
(When a map from a residue ring exists)
Let and let be a nonzero commutative unital ring. A unital homomorphism exists exactly when . Equivalently, has positive characteristic dividing . When it exists, it is unique and sends to . Its image is the prime subring, and it is injective exactly when .
Proof
(Checking the defining relation)
Necessity follows by mapping the relation . Conversely, if , changing by a multiple of changes by zero, so the formula is well-defined. The two displayed identities prove preservation of both operations; the image of one is one. Uniqueness follows because every residue is an integer multiple of . The kernel consists of the residues represented by multiples of . If only the zero residue occurs; if , the nonzero residue is killed.
Proof X-ray
A relation must survive the map
The image of one is already prescribed, so there is no freedom to choose a more convenient image for a generator. The real question is whether the source relation still holds in the target. Checking this relation proves well-definedness; merely writing a representative formula does not. This is the same issue that appears when constructing a map from any quotient.
Worked examples
Worked example 3.1.7
(Solving simultaneous congruences constructively)
Solve , . Choose and , so . The proof gives , and every solution is . Uniqueness modulo fifteen follows because the difference of two solutions lies in . The construction explains both existence and uniqueness.
Worked example 3.1.8
(A domain quotient that is not a field)
Evaluation at zero defines with . Its kernel is , since zero constant term means with . Hence . Thus is prime but not maximal. Prime and maximal ideals coincide in some rings, but not in all rings.
Worked example 3.1.9
(Why the additive subgroup condition is too weak)
The constant polynomials form an additive subgroup containing , but do not form an ideal: multiplying by leaves the constants. The proposed congruence identifies and , yet multiplication by would then have to identify and , which it does not. Thus multiplication cannot descend to the additive quotient by the constants.
Worked example 3.1.10
(A unique unital map can miss most of its target)
Let . Its characteristic is four: a constant polynomial is zero exactly when its coefficient is zero modulo four. There is therefore a unique unital map , sending to the constant polynomial . Its kernel is . Its image has four elements, so it is neither injective nor onto ; the formal polynomial is outside the image.
In the reverse direction there is no unital map . Such a map would send to , which is nonzero. Even an injective additive map need not preserve ring structure: is well-defined and additive, but the image of one is and its square is , rather than .
Counterexample mode
A unit in a quotient need not lift to a unit
The claim that an invertible coset must have an invertible representative is false. The integer two is not a unit in , but its class modulo three is its own inverse because four is congruent to one. The correct condition for to be a unit is : an inverse gives , while an expression with gives an inverse coset. Quotient arithmetic can create units as well as zero divisors.
Common mistakes
Common mistake
(An ideal absorbs every ring element)
In this commutative setting an ideal absorbs multiplication by every element of , not just by its own elements. A subring need not be an ideal, and a proper ideal cannot contain . For noncommutative rings, left, right, and two-sided ideals must be distinguished; the ring-quotient construction uses two-sided ideals. We make no noncommutative generalization silently.
Common mistake
(The Chinese remainder map need not be onto)
The Chinese remainder map always exists, but may fail to be onto without comaximality. For example, a pair of residues modulo four and six must agree modulo two. The pair cannot come from any integer.
Common mistake
(Products of nonzero elements can vanish)
Products of nonzero ring elements may be zero: in , , although neither factor is zero. Cancellation without a domain hypothesis loses solutions.
Summary
Ideals describe exactly the congruences compatible with ring multiplication. Quotients convert ideal properties into familiar arithmetic properties: prime means domain quotient, maximal means field quotient. The Chinese remainder theorem splits arithmetic into independent coordinates only after comaximality is proved. Kernel, surjectivity, and representative checks carry the mathematical content of the construction.
Exercises
Quick check
(Prove that every ideal of is principal. Which are prime, and which are maximal?)
Solution
The zero ideal is . For a nonzero ideal , choose its least positive element (negate a nonzero element if necessary). Division for gives and , so . Thus . The proper prime ideals are and for positive primes : is a domain, and is a field by integer Bézout. For composite with , neither factor belongs to though their product does, so is not prime. Maximal ideals are exactly ; is not maximal since is not a field.
Quick check
(In , find every unit and every nonzero zero divisor, justifying completeness.)
Solution · Answer
A residue is a unit precisely when , by Bézout and the converse divisibility implication from . The units are . The remaining nonzero classes are zero divisors: for , the nonzero class annihilates . These lists cover all nonzero classes.
Quick check
(Show that is a maximal ideal of and is not principal.)
Solution · Answer
The onto map , , has kernel , so the quotient is a field and the ideal maximal. If , then divides the constant polynomial two. Degrees in the domain force to be constant. A constant dividing both two and must divide the coefficient one of , so . That would make the ideal the whole ring, contradicting properness. Hence it is not principal.
Quick check
(Solve , . Explain why this noncoprime system can have solutions even though the Chinese remainder map is not onto.)
Solution · Answer
Write . Modulo six the second condition becomes , or , giving . Thus . The prescribed residues agree modulo , so this particular pair is compatible. Surjectivity would require every pair to be compatible, which fails. The solution period is the least common multiple twelve.
Quick check
(Let . Find its characteristic. Does a unital map exist, and is it injective or surjective? Describe its image and decide whether a unital map exists.)
Solution · Answer
The element vanishes exactly when both four and six divide , or equivalently when twelve divides . Thus the characteristic is twelve. The unique unital map from exists and is injective. Its image, the prime subring, consists of the twelve pairs whose two coordinates have the same parity: necessity follows from a common integer representative, and the substitution argument from the exercise on the noncoprime congruence system, with arbitrary compatible residues, gives sufficiency. It is not surjective, since is missing. No unital map from exists because .