math.wiki
Abstract algebra › Rings in depth › Prime and maximal ideals
Sign in

Prime and maximal ideals

Dividing the integers by n \ge 2 leaves a field when n is prime and a ring with zero divisors otherwise. For an ideal of any commutative ring, what decides whether the quotient is an integral domain, or even a field?

Where you are 4 in Rings in depth

Why this exists

For n \ge 2, the ring \mathbb{Z}_n = \mathbb{Z}/n\mathbb{Z} is a field when n is prime and has zero divisors otherwise (Rings and integral domains, Corollary 8). Two properties of the ideal \langle p \rangle explain this, and both make sense for an ideal of any commutative ring (Ideals and quotient rings). A product lies in \langle p \rangle only if a factor does: p \mid ab forces p \mid a or p \mid b (Every number has exactly one prime factorization). And no ideal fits strictly between \langle p \rangle and \mathbb{Z}. The first property makes a quotient an integral domain, the second makes it a field.

Definition

Throughout, R is a commutative ring with 1. In plain words, an ideal is prime when a product lands in it only if a factor already does, and maximal when it is proper and no ideal fits strictly between it and the whole ring.

  • A prime ideal of R is an ideal P \ne R such that for all a, b \in R, ab \in P implies a \in P or b \in P.
  • A maximal ideal of R is an ideal M \ne R such that the only ideals J with M \subseteq J \subseteq R are J = M and J = R.

A ring homomorphism \psi : R \to S of commutative rings with 1 preserves sums, products and 1; its kernel is \ker \psi = \{a \in R : \psi(a) = 0\} (Ring homomorphisms). Examples are the natural map \pi : R \to R/I (Ideals and quotient rings, Theorem 6), evaluation f \mapsto f(c) on R[x], and composites of these.

For a field F, a polynomial f \in F[x] is irreducible if it is nonconstant and not a product of two polynomials in F[x] of smaller degree (Irreducible polynomials and Eisenstein's criterion). One of degree 2 or 3 is irreducible exactly when it has no root in F: a factorization would have a factor of degree 1, giving a root, and a root c gives the factor x - c (The remainder and factor theorems).

Intuition

Read an ideal as the elements declared to be zero. A prime ideal creates no new zero divisors. A maximal ideal leaves no room: declaring one more element zero forces 1 to be zero, so every element not yet zero can be inverted. In \mathbb{Z}, \langle 6 \rangle fails the first test (2 \cdot 3 lies in it), and \langle 0 \rangle passes the first and fails the second.

\langle 2, x \rangle
\langle 2, x + 1 \rangle
\langle 3, x^2 + 1 \rangle
\langle x \rangle
\langle 2 \rangle
\langle x^2 + 1 \rangle
\langle 3 \rangle
\langle 0 \rangle
Figure 1Some prime ideals of \mathbb{Z}[x] ordered by inclusion. The three on top are maximal; the five below are prime and not maximal (Example 2, Exercises 8 and 9; the ideal generated by 2 as the one generated by 3 in Exercise 8). In \mathbb{Z} a chain of prime ideals has at most two members, \langle 0 \rangle \subsetneq \langle p \rangle; here chains of three occur.

Results

Throughout, R is a commutative ring with 1.

Theorem 1 (prime ideals and domains). An ideal P of R is prime if and only if R/P is an integral domain.

Theorem 2 (ideals of a quotient). Let I be an ideal of R and \pi : R \to R/I the natural map. Then J \mapsto \pi(J) = \{a + I : a \in J\} is a bijection from the ideals of R containing I to the ideals of R/I, with inverse K \mapsto \pi^{-1}(K) = \{a \in R : a + I \in K\}. Both maps preserve inclusion, J = R exactly when \pi(J) = R/I, and J is prime exactly when \pi(J) is prime.

Theorem 3 (maximal ideals and fields). An ideal M of R is maximal if and only if R/M is a field.

Corollary 4. Every maximal ideal of R is prime.

Theorem 5 (kernels). Let \psi : R \to S be a ring homomorphism of commutative rings with 1. (a) If S is an integral domain, then \ker \psi is prime. (b) If \psi is onto and S is a field, then \ker \psi is maximal.

Theorem 6 (the integers). The prime ideals of \mathbb{Z} are \langle 0 \rangle and the \langle p \rangle with p prime; the maximal ideals are the \langle p \rangle. For n \ge 2, the ideals of \mathbb{Z}_n are the \langle [d] \rangle with d a positive divisor of n, and both its prime and its maximal ideals are the \langle [p] \rangle with p a prime dividing n.

Theorem 7 (polynomials over a field). Let F be a field. (a) Every ideal of F[x] is \langle h \rangle for some h \in F[x]. (b) For a nonconstant f \in F[x], \langle f \rangle is maximal if and only if it is prime, if and only if f is irreducible. The ideal \langle 0 \rangle is prime and not maximal. (c) If \deg f = n \ge 1, each coset of \langle f \rangle contains exactly one polynomial of degree less than n (counting 0). So if F has q elements, F[x]/\langle f \rangle has q^n.

Theorem 8 (finite rings). In a finite commutative ring with 1, every prime ideal is maximal.

Why it is true

Proof of Theorem 1. The ring R/P is commutative with identity 1 + P, and 1 + P \ne 0 + P exactly when P \ne R (Ideals and quotient rings, Theorems 2 and 6). Since (a + P)(b + P) = ab + P and a + P = 0 + P exactly when a \in P, “a product of cosets is 0 + P only if a factor is” says “ab \in P only if a \in P or b \in P”. So R/P is an integral domain exactly when P is proper and prime. \square

Proof of Theorem 2. The natural map \pi is onto with kernel I, so the two maps are inverse bijections between these sets of ideals (Ring homomorphisms, Exercise 15). In particular \pi^{-1}(\pi(J)) = J; that is,

a \in J \iff \pi(a) \in \pi(J). \tag{$\ast$}

The two maps are inverse to each other and preserve inclusion. With a = 1, (\ast) gives J = R exactly when \pi(J) = R/I. Every element of R/I is some \pi(a), and \pi(a)\pi(b) = \pi(ab), so by (\ast) the prime condition for \pi(J) is the prime condition for J. \square

Proof of Theorem 3. A commutative ring with 1 \ne 0 is a field if and only if its only ideals are \{0\} and itself (Ideals and quotient rings, Corollary 3); the zero ring has one ideal. So a commutative ring with 1 is a field exactly when it has exactly two ideals. By Theorem 2 with I = M, the ideals of R/M match the ideals of R containing M, with \{0 + M\} matching M and R/M matching R. So R/M is a field exactly when M \ne R and no other ideal contains M. \square

Proof of Corollary 4. R/M is a field (Theorem 3), hence an integral domain (Rings and integral domains, Theorem 5), so M is prime (Theorem 1). \square

Proof of Theorem 5. Let K = \ker \psi, an ideal (Ring homomorphisms, Theorem 3). In both parts 1 \ne 0 in S, so \psi(1) \ne 0 and K \ne R.

(a) If ab \in K, then \psi(a)\psi(b) = 0 in the domain S, so a \in K or b \in K.

(b) Let J \supsetneq K be an ideal and a \in J \setminus K. As \psi(a) \ne 0 and \psi is onto, \psi(b) = \psi(a)^{-1} for some b. Then \psi(ab - 1) = 0, so ab - 1 \in K \subseteq J and 1 = ab - (ab - 1) \in J: J = R. \square

Proof of Theorem 6. Every ideal of \mathbb{Z} is \langle n \rangle for one n \ge 0 (Ideals and quotient rings, Theorem 5), and \langle 1 \rangle is not proper. As \mathbb{Z} is a domain, \langle 0 \rangle is prime, and it is not maximal: \langle 0 \rangle \subsetneq \langle 2 \rangle \subsetneq \mathbb{Z}. For n \ge 2, \mathbb{Z}/\langle n \rangle = \mathbb{Z}_n is a domain if and only if it is a field, if and only if n is prime (Rings and integral domains, Corollary 8); apply Theorems 1 and 3. For \mathbb{Z}_n, apply Theorem 2 with I = n\mathbb{Z}: d\mathbb{Z} \supseteq n\mathbb{Z} means d \mid n, and \pi(d\mathbb{Z}) = \langle [d] \rangle; and as \mathbb{Z}_n is finite, its prime ideals are maximal (Theorem 8). \square

Proof of Theorem 7. For nonzero g, h the product of the leading coefficients is nonzero, so \deg(gh) = \deg g + \deg h. Hence F[x] is an integral domain whose units are the nonzero constants.

(a) Let J \ne \{0\} be an ideal and h \in J nonzero of least degree. For k \in J, Polynomial long division by h (possible since the leading coefficient of h is invertible) gives k = qh + r with r = 0 or \deg r < \deg h. As r = k - qh \in J, r = 0. So J = \langle h \rangle.

(b) Let f be irreducible; it is not a unit, so \langle f \rangle \ne F[x]. An ideal J containing f is \langle h \rangle by (a), with f = gh. If \deg h = 0, h is a unit and J = F[x]; otherwise irreducibility forces \deg g = 0, and J = \langle g^{-1}f \rangle = \langle f \rangle. So \langle f \rangle is maximal, hence prime (Corollary 4). If f is not irreducible, f = gh with 0 < \deg g, \deg h < \deg f; then gh \in \langle f \rangle, while a nonzero multiple of f has degree at least \deg f, so g, h \notin \langle f \rangle. As F[x] is a domain, \langle 0 \rangle is prime; it is not maximal, as \langle 0 \rangle \subsetneq \langle x \rangle \subsetneq F[x].

(c) Division gives k + \langle f \rangle = r + \langle f \rangle with r = 0 or \deg r < n. If two such r, r' give the same coset, then f \mid r - r', of degree less than n, so r = r'. There are q^n choices of the n coefficients. \square

Proof of Theorem 8. R/P is a finite integral domain (Theorem 1), hence a field (Rings and integral domains, Theorem 7), so P is maximal (Theorem 3). \square

Worked examples

Example 1

The prime ideals of \mathbb{Z}_{84}

Find the prime ideals and the maximal ideals of \mathbb{Z}_{84}, with the size of each and of its quotient. Show directly that \langle 6 \rangle and \langle 0 \rangle are not prime.

Solution

Write k for [k]. As 84 = 2^2 \cdot 3 \cdot 7, Theorem 6 gives the prime ideals \langle 2 \rangle, \langle 3 \rangle, \langle 7 \rangle, which are also the maximal ones. As p \mid 84, \langle p \rangle holds the multiples of p from 0 to 83, so it has 84/p elements: \langle 2 \rangle has 42, \langle 3 \rangle has 28, \langle 7 \rangle has 12. A coset a + \langle p \rangle is fixed by a \bmod p, so the quotients have 2, 3 and 7 elements, and each is a field by Theorem 3.

Directly: 2 \cdot 3 = 6 \in \langle 6 \rangle, but 2 and 3 are not among the multiples 0, 6, \dots, 78 of 6, so \langle 6 \rangle is not prime. And 12 \cdot 7 = 84 = 0 with 12 \ne 0 and 7 \ne 0, so \langle 0 \rangle is not prime.

Example 2

Prime but not maximal in \mathbb{Z}[x]

In \mathbb{Z}[x], show that \langle x \rangle is prime but not maximal, and that \langle 2, x \rangle is maximal.

Solution

Since f = f(0) + xg with g \in \mathbb{Z}[x], f \in \langle x \rangle exactly when f(0) = 0. Evaluation \varepsilon(f) = f(0) is a ring homomorphism from \mathbb{Z}[x] to the integral domain \mathbb{Z} with kernel \langle x \rangle, so \langle x \rangle is prime by Theorem 5(a).

Every 2g + xh has even constant term 2g(0), so 1 \notin \langle 2, x \rangle, while 2 \notin \langle x \rangle. So \langle x \rangle \subsetneq \langle 2, x \rangle \subsetneq \mathbb{Z}[x], and \langle x \rangle is not maximal.

Let \psi(f) = [f(0)]_2, the composite of \varepsilon and the natural map onto the field \mathbb{Z}_2. If f(0) = 2k, then f = 2k + xg \in \langle 2, x \rangle; with the previous paragraph, \ker \psi = \langle 2, x \rangle, which is maximal by Theorem 5(b).

Example 3

A field with 27 elements

Let f = x^3 + 2x + 1 in \mathbb{Z}_3[x] and I = \langle f \rangle, writing k for [k]. Show that \mathbb{Z}_3[x]/I is a field with 27 elements and find the inverse of x + I. Is \langle x^3 + 2 \rangle prime?

Solution

We have f(0) = 1, f(1) = 4 = 1 and f(2) = 13 = 1, so the cubic f has no root in \mathbb{Z}_3 and is irreducible. By Theorems 7(b) and 3, \mathbb{Z}_3[x]/I is a field, and by Theorem 7(c) its elements are the 3^3 = 27 cosets a + bx + cx^2 + I.

In the quotient x^3 + I = -2x - 1 + I = x + 2 + I, so x^3 + 2x reduces to (x + 2) + 2x = 3x + 2 = 2. As 2 \cdot 2 = 1 in \mathbb{Z}_3, x \cdot 2(x^2 + 2) \equiv 1, and 2(x^2 + 2) = 2x^2 + 1. So (x + I)^{-1} = 2x^2 + 1 + I. Check: 2x^3 + x reduces to 2(x + 2) + x = 3x + 4 = 1.

No: (x + 2)^3 = x^3 + 6x^2 + 12x + 8 = x^3 + 2 in \mathbb{Z}_3[x], so (x + 2)(x + 2)^2 lies in \langle x^3 + 2 \rangle while neither factor does (their degrees are below 3).

Example 4

Gaussian integers modulo 3 and modulo 5

In \mathbb{Z}[i], show that \langle 3 \rangle is maximal and find the inverse of 1 + i + \langle 3 \rangle. Show that \langle 5 \rangle is not prime.

Solution

The elements of \langle 3 \rangle are 3(c + di) = 3c + 3di, so a + bi \in \langle 3 \rangle exactly when 3 \mid a and 3 \mid b. So the cosets are fixed by a \bmod 3 and b \bmod 3: there are 9, and 1 \notin \langle 3 \rangle.

A square is \equiv 0 or 1 \pmod 3, so 3 \mid a^2 + b^2 only when 3 \mid a and 3 \mid b. Hence \alpha \in \langle 3 \rangle exactly when 3 \mid N(\alpha). If \alpha\beta \in \langle 3 \rangle, then 3 divides N(\alpha\beta) = N(\alpha)N(\beta) (as N(z) = z\bar z, Complex numbers), so, 3 being prime, it divides N(\alpha) or N(\beta). Thus \langle 3 \rangle is prime, and as the quotient is finite, maximal by Theorem 8: \mathbb{Z}[i]/\langle 3 \rangle is a field with 9 elements. As (1 + i)(2 + i) = 1 + 3i \equiv 1, the inverse is 2 + i + \langle 3 \rangle.

For 5: (2 + i)(2 - i) = 4 - i^2 = 5 \in \langle 5 \rangle, while 2 \pm i \notin \langle 5 \rangle, so \langle 5 \rangle is not prime. Here 5 = 2^2 + 1^2 is a sum of two squares; 3 is not.

Example 5

The prime ideals of a product

Let R and S be commutative rings with 1. Prove that the prime ideals of R \times S are exactly the sets P \times S with P prime in R and R \times Q with Q prime in S.

Solution

Proof. Let P' be prime in R \times S. As (1, 0)(0, 1) = (0, 0), one of (1, 0), (0, 1) lies in P'; not both, or (1, 1) \in P'. Say (0, 1) \in P'. Then (0, s) = (0, s)(0, 1) \in P' for all s, so (a, b) \in P' exactly when (a, 0) = (a, b) - (0, b) \in P'. Let P = \{a \in R : (a, 0) \in P'\}, so P' = P \times S. From (a, 0) - (a', 0) = (a - a', 0) and (r, 0)(a, 0) = (ra, 0), P is an ideal; 1 \notin P as (1, 0) \notin P'; and if ab \in P, then (a, 0)(b, 0) \in P', so a \in P or b \in P. The case (1, 0) \in P' gives R \times Q in the same way.

Conversely, for P prime, (a, b) \mapsto a + P is a ring homomorphism from R \times S into the integral domain R/P (Theorem 1) with kernel P \times S, which is prime by Theorem 5(a); likewise R \times Q. \square

Boundaries

  • Commutativity. In M_2(\mathbb{R}) the only ideals are \{0\} and the whole ring (Ideals and quotient rings, Exercise 17), so \{0\} is maximal; yet M_2(\mathbb{R}) is not a field, and E_{11}E_{22} = 0 with neither factor 0, so \{0\} fails the element test for prime. Theorems 1 and 3 and Corollary 4 need R commutative.
  • Prime is weaker than maximal. \langle 0 \rangle in a domain that is not a field, and the figure’s middle row, are prime and not maximal. In \mathbb{Z} and F[x] every nonzero prime ideal is maximal; Euclidean domains and principal ideal domains extends this to every principal ideal domain.
  • Existence. Every proper ideal of a commutative ring with 1 lies in a maximal ideal (Krull’s theorem); the general proof uses Zorn’s lemma and is beyond this lesson. For finite rings see Exercise 15.

Where this goes next

For an irreducible f \in F[x], the field F[x]/\langle f \rangle contains a copy of F (the constant cosets), and in it f(x + \langle f \rangle) = f + \langle f \rangle = 0: the polynomial has acquired a root. This is how Field extensions and degree builds larger fields; Examples 3 and 4 are first cases of Finite fields. Euclidean domains and principal ideal domains carries Theorem 7(a) to every ring with a division algorithm, and Unique factorization domains explains why 5 splits in \mathbb{Z}[i] and 3 does not.

Exercises

Ideals of \mathbb{Z}_n

  1. List the prime ideals and the maximal ideals of \mathbb{Z}_{90}, with the number of elements of each ideal and of its quotient.
  2. Yuna and Mateo play a parlor game on \mathbb{Z}_{60}. Yuna picks an ideal I; Mateo wins if he can name a, b \notin I with ab \in I. For how many choices of I can Mateo win, and what pair can he name against \langle 12 \rangle?

Quotients of polynomial rings

  1. In \mathbb{Z}_5[x], decide whether \langle f \rangle is maximal, and if so give the number of elements of the quotient: (a) f = x^2 + 2; (b) f = x^2 + 4; (c) f = x^3 + x + 1.
  2. Let I = \langle x^2 + 2 \rangle in \mathbb{Z}_5[x]. Find the inverse of x + 1 + I.
  3. Show that I = \langle x^2 - 6x + 7 \rangle is maximal in \mathbb{Q}[x], and find the inverse of x - 3 + I.
  4. Show that \langle x^4 - 4 \rangle is not prime in \mathbb{Q}[x], and find every maximal ideal of \mathbb{Q}[x] containing it.

Kernels and other rings

  1. Let \psi : \mathbb{Z}[x] \to \mathbb{Z}_7, \psi(f) = [f(3)]_7. Show that \ker \psi = \langle 7, x - 3 \rangle, that this ideal is maximal, and write x^2 + 5 as an element of it.
  2. In \mathbb{Z}[x], show that \langle 3 \rangle is prime but not maximal and that \langle 2, x + 1 \rangle is maximal.
  3. In \mathbb{Z}[x], show that \langle x^2 + 1 \rangle is prime and that \langle 3, x^2 + 1 \rangle is maximal. (Use evaluation at i and Example 4.)
  4. In \mathbb{Z}[i], is \langle 7 \rangle maximal? Is \langle 13 \rangle prime?

Proofs

  1. Let P be prime and a^n \in P for a positive integer n. Prove a \in P, and deduce that every prime ideal contains every nilpotent element.
  2. Let P be prime and I, J ideals with I \cap J \subseteq P. Prove that I \subseteq P or J \subseteq P.
  3. Let \psi : R \to S be a ring homomorphism of commutative rings and Q a prime ideal of S. Prove that \psi^{-1}(Q) is a prime ideal of R. Show by an example that the preimage of a maximal ideal need not be maximal.
  4. Let R be a commutative ring with 1. Prove that \langle x \rangle is prime in R[x] if and only if R is an integral domain, and maximal if and only if R is a field.
  5. Prove that every proper ideal of a finite commutative ring with 1 lies in a maximal ideal.
Answers
  1. 90 = 2 \cdot 3^2 \cdot 5, so by Theorem 6 the prime ideals, also the maximal ones, are \langle 2 \rangle (45 elements), \langle 3 \rangle (30) and \langle 5 \rangle (18); the quotients have 2, 3 and 5 elements.
  2. Mateo wins exactly when I is proper and not prime. Of the 12 ideals (one per divisor of 60), one is not proper and three are prime (\langle 2 \rangle, \langle 3 \rangle, \langle 5 \rangle), so he wins for 8 choices. Against \langle 12 \rangle = \{0, 12, 24, 36, 48\} he names 3 and 4, as 3 \cdot 4 = 12.
  3. (a) Maximal, with 5^2 = 25 elements: the squares in \mathbb{Z}_5 are 0, 1, 4, and -2 = 3 is not one, so f has no root. (b) Not maximal: x^2 + 4 = (x + 1)(x + 4), as 5x = 0. (c) Maximal, with 5^3 = 125 elements: f(0), \dots, f(4) are 1, 3, 1, 1, 4, so the cubic has no root.
  4. 3x + 2 + I: (x + 1)(3x + 2) = 3x^2 + 5x + 2 = 3x^2 + 2, and x^2 + I = 3 + I, giving 9 + 2 = 11 = 1.
  5. By the The rational root theorem a rational root would be \pm 1 or \pm 7, where the values are 2, 14, 14, 98. A quadratic with no root is irreducible, so I is maximal (Theorem 7). As (x - 3)^2 = (x^2 - 6x + 7) + 2, the inverse is \displaystyle \frac{1}{2}x - \frac{3}{2} + I.
  6. x^4 - 4 = (x^2 - 2)(x^2 + 2), with neither factor in the ideal (degree 2). A maximal M containing it is prime, so contains x^2 - 2 or x^2 + 2; both are irreducible (no rational roots), so generate maximal ideals, which must equal M. The answer: \langle x^2 - 2 \rangle and \langle x^2 + 2 \rangle.
  7. \psi is onto, as \psi(c) = [c]_7. Dividing by the monic x - 3 gives f = (x - 3)g + f(3) in \mathbb{Z}[x], so \psi(f) = 0 exactly when f(3) = 7k, exactly when f \in \langle 7, x - 3 \rangle. Maximal by Theorem 5(b), and x^2 + 5 = (x + 3)(x - 3) + 2 \cdot 7.
  8. Reducing coefficients mod 3 maps \mathbb{Z}[x] onto the domain \mathbb{Z}_3[x] with kernel \langle 3 \rangle, so it is prime; not maximal, as \langle 3 \rangle \subsetneq \langle 3, x \rangle \subsetneq \mathbb{Z}[x] (constant terms of 3g + xh are multiples of 3). For \langle 2, x + 1 \rangle, take f \mapsto [f(-1)]_2, onto \mathbb{Z}_2; as f = (x + 1)g + f(-1), its kernel is \langle 2, x + 1 \rangle, maximal by Theorem 5(b).
  9. Evaluation f \mapsto f(i) maps into the domain \mathbb{C}. Dividing by the monic x^2 + 1 gives f = q(x^2 + 1) + a + bx with a, b \in \mathbb{Z}, so f(i) = a + bi, and the kernel is \langle x^2 + 1 \rangle: prime by Theorem 5(a). Then f \mapsto f(i) + \langle 3 \rangle maps onto the field of Example 4, with kernel the f having 3 \mid a, 3 \mid b, namely \langle 3, x^2 + 1 \rangle: maximal by Theorem 5(b).
  10. Yes. The squares mod 7 are 0, 1, 2, 4, and no two of them add to 0 or 7 except 0 + 0, so 7 \mid a^2 + b^2 forces 7 \mid a and 7 \mid b. As in Example 4, \langle 7 \rangle is prime, and the quotient has 49 elements, so it is maximal (Theorem 8). No for 13: (3 + 2i)(3 - 2i) = 13, and 13 \nmid 3.
  11. Proof. By induction on n; n = 1 is clear. If a \cdot a^{n - 1} \in P with n \ge 2, then a \in P or a^{n - 1} \in P, and in the second case a \in P by the induction hypothesis. A nilpotent a has a^n = 0 \in P for some n, so a \in P. \square
  12. Proof. Suppose a \in I \setminus P and b \in J \setminus P. Then ab \in I and ab \in J, as ideals absorb, so ab \in I \cap J \subseteq P, and primality puts a or b in P, a contradiction. \square
  13. Proof. Let P = \psi^{-1}(Q), an ideal of R (Ring homomorphisms, Theorem 2(b)), proper as \psi(1) = 1 \notin Q. If ab \in P, then \psi(a)\psi(b) \in Q, so \psi(a) or \psi(b) lies in Q. \square Example: the inclusion \mathbb{Z} \to \mathbb{Q} pulls the maximal ideal \{0\} of \mathbb{Q} back to \langle 0 \rangle, not maximal in \mathbb{Z}.
  14. Proof. Evaluation f \mapsto f(0) maps R[x] onto R with kernel \langle x \rangle, as f = f(0) + xg; so Theorem 5 gives one direction of each claim. Conversely let \langle x \rangle be prime. Being proper, it misses 1, so 1 \ne 0 in R. Its only constant is 0, and ab = 0 in R puts a or b in \langle x \rangle, making it 0: R is a domain. If \langle x \rangle is maximal and a \ne 0, then \langle x \rangle + \langle a \rangle (Ideals and quotient rings, Exercise 13) properly contains \langle x \rangle, so 1 = xg + ah; comparing constant terms, 1 = a\,h(0), and a is a unit. So R is a field. \square
  15. Proof. Let I be proper in the finite ring R. Among the proper ideals containing I (finitely many, I among them) choose M with the most elements. If J is an ideal with M \subseteq J \ne R, then J is in the set, so |J| \le |M|, forcing J = M. So M is maximal and contains I. \square
math.wiki · updated 2026-11-17 · CC-BY-SA-4.0