math.wiki
Abstract algebra › Rings › Ring homomorphisms
Sign in

Ring homomorphisms

A group homomorphism respects one operation, and its kernels are exactly the normal subgroups. What must a map between rings respect, and which subsets of a ring are its kernels?

Where you are 4 in Rings

Why this exists

A group homomorphism (Homomorphisms and kernels) respects the one operation of a group; its kernel is a normal subgroup, and every normal subgroup is a kernel. A ring has two operations. Reduction mod n respects both, so a question about \mathbb{Z} can be carried to the finite ring \mathbb{Z}_n and settled by checking cases. Maps that respect sums, products and 1 are the ring homomorphisms. Their kernels are exactly the Ideals and quotient rings, and the first isomorphism theorem identifies a quotient ring R/I with a ring already known.

Definition

Let R and S be rings, and write 1_S for the identity of S. A ring homomorphism is a map \psi : R \to S such that, for all a, b \in R,

\begin{aligned} \psi(a + b) &= \psi(a) + \psi(b), \\ \psi(ab) &= \psi(a)\psi(b), \\ \psi(1) &= 1_S. \end{aligned}

Its kernel is \ker \psi = \{a \in R : \psi(a) = 0\} and its image is \operatorname{im} \psi = \{\psi(a) : a \in R\}. A ring isomorphism is a ring homomorphism that is a bijection (One-to-one, onto and bijective functions); R \cong S means one exists. A ring isomorphism R \to R is an automorphism.

The condition \psi(1) = 1_S belongs to the definition here because rings have an identity (Rings and integral domains); Judson and Beezer omit it (see Boundaries). For an integer k, k \cdot 1_S is the integer multiple of 1_S (the sum of k copies of 1_S when k > 0).

Intuition

A ring homomorphism is a translation under which every true calculation with + and \times stays true: reduction mod 9 turns 12^3 + 7^3 = 2071 into 3^3 + 7^3 = 1, true in \mathbb{Z}_9. The kernel is what the translation forgets. Since 0 \cdot s = 0, forgetting a forces forgetting ra, so the kernel is an ideal. The first isomorphism theorem says nothing else is lost: R / \ker \psi is a copy of the image.

Results

Throughout, \psi : R \to S is a ring homomorphism and K = \ker \psi.

Theorem 1 (what a ring homomorphism keeps). (a) \psi(0) = 0, \psi(-a) = -\psi(a), \psi(ka) = k\psi(a) for every integer k, and \psi(a^n) = \psi(a)^n for n \ge 0. (b) If R and S are commutative and f(x_1, \dots, x_n) is a polynomial with integer coefficients, then \psi(f(a_1, \dots, a_n)) is f(\psi(a_1), \dots, \psi(a_n)). (c) If u \in R^\times, then \psi(u) \in S^\times and \psi(u^{-1}) = \psi(u)^{-1}; so \psi restricts to a group homomorphism R^\times \to S^\times. (d) If \theta : S \to T is a ring homomorphism, so is \theta \circ \psi. The inverse of a ring isomorphism is a ring isomorphism.

Theorem 2 (images and preimages). (a) \operatorname{im} \psi is a subring of S. (b) If J is an ideal of S, then \psi^{-1}(J) is an ideal of R containing K. (c) If \psi is onto and I is an ideal of R, then \psi(I) is an ideal of S.

Theorem 3 (kernels are the ideals). K is an ideal of R, and \psi is one to one if and only if K = \{0\}. Conversely, every ideal I of R is the kernel of a ring homomorphism, the natural map \pi : R \to R/I.

Corollary 4 (maps out of a field). If F is a field and S is not the zero ring, every ring homomorphism F \to S is one to one.

Theorem 5 (maps out of the integers). Let S be a ring. (a) The map k \mapsto k \cdot 1_S is the only ring homomorphism \mathbb{Z} \to S, and its kernel is (\operatorname{char} S)\mathbb{Z}. (b) Let n \ge 1. There is a ring homomorphism \mathbb{Z}_n \to S if and only if n \cdot 1_S = 0; it is then unique, [k] \mapsto k \cdot 1_S. In particular a ring homomorphism \mathbb{Z}_m \to \mathbb{Z}_n exists if and only if n \mid m, and it is [k]_m \mapsto [k]_n.

Theorem 6 (first isomorphism theorem for rings). The map \theta : R/K \to \operatorname{im} \psi, \theta(a + K) = \psi(a), is a well-defined ring isomorphism. So R/\ker \psi \cong \operatorname{im} \psi, and R/\ker \psi \cong S if \psi is onto.

Corollary 7 (the prime subring). \{k \cdot 1_R : k \in \mathbb{Z}\} is a subring of R contained in every subring of R. It is isomorphic to \mathbb{Z}_n if \operatorname{char} R = n > 0, and to \mathbb{Z} if \operatorname{char} R = 0.

Why it is true

Proof of Theorem 1. (a) \psi is a homomorphism of additive groups, so the first three statements are Homomorphisms and kernels, Theorem 1, written additively. Next \psi(a^0) = \psi(1) = 1_S = \psi(a)^0, and \psi(a^{n + 1}) = \psi(a^n)\psi(a), so induction gives \psi(a^n) = \psi(a)^n. (b) f(a_1, \dots, a_n) is a sum of terms k\, a_1^{e_1} \cdots a_n^{e_n}; by (a) and the product rule, \psi sends each term to k\, \psi(a_1)^{e_1} \cdots \psi(a_n)^{e_n}, and it respects the sum. (c) \psi(u)\psi(u^{-1}) = \psi(uu^{-1}) = \psi(1) = 1_S, and likewise in the other order. (d) \theta(\psi(ab)) = \theta(\psi(a)\psi(b)), which is \theta(\psi(a))\theta(\psi(b)); the same holds for sums, and \theta(\psi(1)) = \theta(1_S) = 1_T. If \psi is a bijection and x = \psi(a), y = \psi(b), then \psi(ab) = xy, so \psi^{-1}(xy) = ab = \psi^{-1}(x)\psi^{-1}(y); sums are the same, and \psi^{-1}(1_S) = 1. \square

Proof of Theorem 2. (a) 1_S = \psi(1), \psi(a) - \psi(b) = \psi(a - b) and \psi(a)\psi(b) = \psi(ab) lie in the image; apply the subring test (Rings and integral domains, Theorem 3). (b) J is a subgroup of (S, +), so \psi^{-1}(J) is a subgroup of (R, +) containing K (Homomorphisms and kernels, Theorem 3(b)). If r \in R and \psi(a) \in J, then \psi(ra) = \psi(r)\psi(a) and \psi(ar) = \psi(a)\psi(r) lie in J, as J absorbs multiplication. (c) \psi(I) is a subgroup of (S, +) (Homomorphisms and kernels, Theorem 3(a)). Let s \in S and a \in I. As \psi is onto, s = \psi(r) for some r, and s\psi(a) = \psi(ra), \psi(a)s = \psi(ar) lie in \psi(I). \square

Proof of Theorem 3. \{0\} is an ideal of S, so K = \psi^{-1}(\{0\}) is an ideal by Theorem 2(b). The criterion for one to one is Homomorphisms and kernels, Corollary 6, for the additive groups. For the converse, Ideals and quotient rings, Theorem 6(c), shows \pi respects sums and products, sends 1 to the identity 1 + I of R/I, and has \pi(a) = 0 + I exactly when a \in I. \square

Proof of Corollary 4. A field is commutative with 1 \ne 0, so its only ideals are \{0\} and F (Ideals and quotient rings, Corollary 3). If K = F, then 1_S = \psi(1) = 0 and S is the zero ring (Rings and integral domains, Theorem 1). So K = \{0\}, and Theorem 3 applies. \square

Proof of Theorem 5. (a) A ring homomorphism \mathbb{Z} \to S is a homomorphism of additive groups sending 1 to 1_S, and by Homomorphisms and kernels, Theorem 9(a), the only one is k \mapsto k \cdot 1_S. This map respects products, since (j \cdot 1_S)(k \cdot 1_S) = (jk) \cdot 1_S by the distributive laws and the rules of signs (Rings and integral domains, Theorem 1). Its kernel is a subgroup of \mathbb{Z}, so it is m\mathbb{Z} for one m \ge 0 (Cyclic groups, Corollary 7), where m is the least positive k with k \cdot 1_S = 0, or 0 if there is none: by definition, m = \operatorname{char} S.

(b) Let \psi : \mathbb{Z}_n \to S be a ring homomorphism. Composed with the natural map k \mapsto [k] it is a ring homomorphism \mathbb{Z} \to S (Theorem 1(d)), so by (a) \psi([k]) = k \cdot 1_S; in particular n \cdot 1_S = \psi([n]) = \psi([0]) = 0. Conversely, if n \cdot 1_S = 0 and [j] = [k], then n \mid j - k and j \cdot 1_S - k \cdot 1_S = (j - k) \cdot 1_S = 0, so [k] \mapsto k \cdot 1_S is well defined. It respects sums, products and 1 because k \mapsto k \cdot 1_S does and [j] + [k] = [j + k], [j][k] = [jk]. For S = \mathbb{Z}_n and n replaced by m, the condition is [m]_n = 0, that is, n \mid m. \square

Proof of Theorem 6. By Homomorphisms and kernels, Theorem 5, for the additive groups, \psi(a) = \psi(b) if and only if a - b \in K, that is, a + K = b + K. Read left to right this makes \theta well defined, and right to left one to one; it is onto \operatorname{im} \psi by definition. With the operations of R/K (Ideals and quotient rings, Definition), \theta\bigl((a + K)(b + K)\bigr) = \theta(ab + K), which is \psi(ab) = \psi(a)\psi(b); sums are the same, and \theta(1 + K) = \psi(1) = 1_S. The group version is The isomorphism theorems. \square

Proof of Corollary 7. The set is the image of k \mapsto k \cdot 1_R, a subring by Theorem 2(a), with kernel n\mathbb{Z}, n = \operatorname{char} R (Theorem 5(a)). If n > 0, Theorem 6 gives an isomorphism from \mathbb{Z}/n\mathbb{Z} = \mathbb{Z}_n onto it. If n = 0, the map is one to one (Theorem 3), so an isomorphism onto its image. A subring contains 1_R and is closed under sums and negatives, so it contains every k \cdot 1_R. \square

Worked examples

Example 1

Three maps to test

Decide which are ring homomorphisms: (a) complex conjugation z \mapsto \bar z from \mathbb{C} to \mathbb{C}; (b) \det : M_2(\mathbb{R}) \to \mathbb{R}; (c) \varepsilon : \mathbb{R} \to M_2(\mathbb{R}), \varepsilon(a) = \begin{pmatrix} a & 0 \\ 0 & 0 \end{pmatrix}.

Solution

(a) Let z = a + bi and w = c + di. Conjugation respects sums coordinate by coordinate. Also zw = (ac - bd) + (ad + bc)i, while

\begin{aligned} \bar z \bar w &= (a - bi)(c - di) \\ &= (ac - bd) - (ad + bc)i = \overline{zw}, \end{aligned}

and \bar 1 = 1. Conjugating twice gives z back, so the map is a bijection: a ring automorphism of \mathbb{C}.

(b) \det(AB) = \det A \det B and \det I_2 = 1 (Properties of determinants: row operations, products, transposes), but \det(I_2 + I_2) = \det(2I_2) = 4, while \det I_2 + \det I_2 = 2. Not a ring homomorphism: products and the identity are not enough.

(c) Sums and products of diagonal matrices are taken entry by entry, so \varepsilon(a + b) = \varepsilon(a) + \varepsilon(b) and \varepsilon(ab) = \varepsilon(a)\varepsilon(b). But \varepsilon(1) \ne I_2, so \varepsilon is not a ring homomorphism under this definition.

Example 2

The integers mod 30 and mod 6

Find every ring homomorphism \mathbb{Z}_{30} \to \mathbb{Z}_6 and \mathbb{Z}_6 \to \mathbb{Z}_{30}, and use the first to show \mathbb{Z}_{30}/\langle 6 \rangle \cong \mathbb{Z}_6.

Solution

By Theorem 5(b) there is exactly one ring homomorphism \psi : \mathbb{Z}_{30} \to \mathbb{Z}_6, as 6 \mid 30: \psi([k]_{30}) = [k]_6. It is onto. Writing k for [k]_{30}, \psi(k) = 0 exactly when 6 \mid k, so \ker \psi = \{0, 6, 12, 18, 24\} = \langle 6 \rangle. By Theorem 6, k + \langle 6 \rangle \mapsto [k]_6 is an isomorphism \mathbb{Z}_{30}/\langle 6 \rangle \to \mathbb{Z}_6. This makes “the same tables” of Ideals and quotient rings, Example 2, exact.

As 30 \nmid 6, there is no ring homomorphism \mathbb{Z}_6 \to \mathbb{Z}_{30}: it would send [1]_6 to [1]_{30}, and 0 = 6[1]_6 to 6[1]_{30} = [6]_{30} \ne 0. As additive groups there are \gcd(6, 30) = 6 homomorphisms (Homomorphisms and kernels, Exercise 17), [k]_6 \mapsto [hk]_{30} for h = 0, 5, 10, 15, 20, 25; none sends [1]_6 to [1]_{30}.

Example 3

Hiro’s crossword

Hiro is building a number crossword, and Amara tests his clues. One clue asks for a sum of the cubes of two integers. Show that 2{,}071 is such a sum and that 2{,}029 is not.

Solution

12^3 + 7^3 = 1728 + 343 = 2071. Now suppose x^3 + y^3 = 2029 with integers x, y, and apply the reduction map \psi : \mathbb{Z} \to \mathbb{Z}_9 of Theorem 5(a). By Theorem 1(b), \psi(x)^3 + \psi(y)^3 = \psi(2029) in \mathbb{Z}_9. Writing k for [k], the cubes of 0, 1, \dots, 8 are

0,\ 1,\ 8,\ 0,\ 1,\ 8,\ 0,\ 1,\ 8,

since 27, 216 are \equiv 0, 64, 343 are \equiv 1 and 125, 512 are \equiv 8. So a sum of two cubes in \mathbb{Z}_9 is one of 0 + 0, 0 + 1, 0 + 8, 1 + 1, 1 + 8, 8 + 8, that is, 0, 1, 8, 2, 0, 7. But 2029 = 225 \cdot 9 + 4, so \psi(2029) = 4, which is not on the list. No such x, y exist. The test can only rule out: passing it proves nothing.

Example 4

Real polynomials modulo x^2 + 1, exactly

Let \psi : \mathbb{R}[x] \to \mathbb{C}, \psi(f) = f(i). Show that \psi is an onto ring homomorphism with kernel I = \langle x^2 + 1 \rangle, so \mathbb{R}[x]/I \cong \mathbb{C}, and find the representative a + bx of the coset of x^3 + 2x^2 + 5.

Solution

Substituting a number respects sums and products, (f + g)(i) = f(i) + g(i) and (fg)(i) = f(i)g(i), because polynomials are multiplied by the distributive law and \mathbb{C} is commutative (Polynomial rings and the division algorithm proves this in general); the constant 1 goes to 1. As \psi(a + bx) = a + bi, \psi is onto. Since i^2 + 1 = 0, the kernel, an ideal (Theorem 3), contains x^2 + 1 and so contains I (Ideals and quotient rings, Theorem 4). Conversely, Polynomial long division gives f = q(x^2 + 1) + (a + bx) with a, b real, so f(i) = a + bi; if this is 0, then a = b = 0 and f \in I. Theorem 6 now gives the isomorphism \theta(f + I) = f(i) onto \mathbb{C}, the “copy of \mathbb{C}” of Ideals and quotient rings, Example 5.

For f = x^3 + 2x^2 + 5, f(i) = -i - 2 + 5 = 3 - i = \theta(3 - x + I). As \theta is one to one, f + I = 3 - x + I; indeed f = (x + 2)(x^2 + 1) + (3 - x).

Example 5

The Gaussian integers modulo 2 - i

Show that \psi(a + bi) = [a + 2b]_5 is a ring homomorphism from \mathbb{Z}[i] onto \mathbb{Z}_5 with kernel \langle 2 - i \rangle, and conclude \mathbb{Z}[i]/\langle 2 - i \rangle \cong \mathbb{Z}_5.

Solution

The idea: 2^2 = 4 = -1 in \mathbb{Z}_5, so [2] can play the part of i. Sums are respected because a + 2b is additive in (a, b). For products, (a + bi)(c + di) = (ac - bd) + (ad + bc)i maps to [ac - bd + 2ad + 2bc], while [a + 2b][c + 2d] = [ac + 2ad + 2bc + 4bd]; the two differ by [5bd] = 0. Also \psi(1) = [1], and \psi(k) = [k] for integers k, so \psi is onto.

\psi(2 - i) = [2 - 2] = 0, so \langle 2 - i \rangle \subseteq \ker \psi. Conversely, if 5 \mid a + 2b, write

a + bi = (a + 2b) - b(2 - i).

Here a + 2b = 5m = (2 - i)(2 + i)m, so both terms lie in \langle 2 - i \rangle. Hence \ker \psi = \langle 2 - i \rangle, and Theorem 6 gives \mathbb{Z}[i]/\langle 2 - i \rangle \cong \mathbb{Z}_5, a field: every Gaussian integer is congruent modulo 2 - i to exactly one of 0, 1, 2, 3, 4.

1 3 0 2 4 1 3 2 4 1 3 0 2 4 3 0 2 4 1 3 0 4 1 3 0 2 4 1 0 2 4 1 3 0 2 1 3 0 2 4 1 3 2 4 1 3 0 2 4
Figure 1The map a + bi \mapsto [a + 2b]_5 of Example 5 on the Gaussian integers with |a|, |b| \le 3, origin at the center. The boxed points are the kernel \langle 2 - i \rangle, a tilted square grid spanned by 2 - i and i(2 - i) = 1 + 2i; each other digit marks a coset.
Example 6

Two rings that differ

For d = 2, 3 let \displaystyle \mathbb{Z}[\sqrt{d}] = \{a + b\sqrt{d} : a, b \in \mathbb{Z}\}, a subring of \mathbb{R}. Prove that there is no ring homomorphism \displaystyle \mathbb{Z}[\sqrt{2}] \to \mathbb{Z}[\sqrt{3}], although \displaystyle a + b\sqrt{2} \mapsto a + b\sqrt{3} is an isomorphism of additive groups.

Solution

Proof. Since \displaystyle \sqrt{d} is irrational, \displaystyle a + b\sqrt{d} = a' + b'\sqrt{d} forces b = b' and then a = a'. So the given map is a well-defined bijection, and it respects sums.

Suppose \psi is a ring homomorphism, and let \displaystyle \psi(\sqrt{2}) = a + b\sqrt{3}. By Theorem 1(a), \psi(2) = 2, so

\begin{aligned} 2 = \psi(\sqrt{2})^2 &= (a + b\sqrt{3})^2 \\ &= (a^2 + 3b^2) + 2ab\sqrt{3}. \end{aligned}

By the uniqueness above, ab = 0 and a^2 + 3b^2 = 2. If b = 0, then a^2 = 2; if a = 0, then 3b^2 = 2. Neither has an integer solution. \square

Boundaries

  • The identity condition. Without \psi(1) = 1_S, the zero map, \varepsilon of Example 1 and [k]_6 \mapsto [25k]_{30} (as 25^2 \equiv 25, 6 \cdot 25 \equiv 0 \pmod{30}) count, and Theorem 5 fails. An onto map respecting sums and products does send 1 to 1_S, since \psi(1)\psi(r) = \psi(r) = \psi(r)\psi(1).
  • Images of ideals. Theorem 2(c) needs \psi onto. The inclusion \mathbb{Z} \to \mathbb{Q} sends the ideal 2\mathbb{Z} to a set that holds 2 but not \displaystyle \frac{1}{2} \cdot 2 = 1.
  • Kernel and image differ in kind. A proper kernel misses 1, so it is an ideal but not a subring; the image is a subring and rarely an ideal.
  • Group maps are not ring maps. An additive isomorphism need not respect products (Example 6).

Where this goes next

Polynomial rings and the division algorithm proves that evaluation is a ring homomorphism out of F[x] and describes the quotients F[x]/\langle f \rangle as in Example 4. Prime and maximal ideals shows a kernel is maximal when the image is a field, as for \langle 2 - i \rangle in Example 5. By Corollary 4 a field maps into a ring only as a copy of itself, which is how Field extensions and degree places one field inside another; the Frobenius map of Exercise 16 returns in Finite fields.

Exercises

Testing maps

  1. Which are ring homomorphisms? (a) k \mapsto -k, \mathbb{Z} \to \mathbb{Z}; (b) a \mapsto aI_2, \mathbb{R} \to M_2(\mathbb{R}); (c) (a, b) \mapsto a + b, \mathbb{Z} \times \mathbb{Z} \to \mathbb{Z}; (d) the trace, M_2(\mathbb{R}) \to \mathbb{R}.
  2. Decide whether there is a ring homomorphism, and if so give it with its kernel: \mathbb{Z}_{12} \to \mathbb{Z}_4, \mathbb{Z}_4 \to \mathbb{Z}_{12}, \mathbb{Z}_{15} \to \mathbb{Z}_5, \mathbb{Z} \to \mathbb{Z}_7, \mathbb{Z}_7 \to \mathbb{Z}.
  3. Find the characteristic and the prime subring (Corollary 7) of \mathbb{Z}_4 \times \mathbb{Z}_6 and of \mathbb{Z}_5 \times \mathbb{Q}.
  4. Find every ring homomorphism \mathbb{Z} \times \mathbb{Z} \to \mathbb{Z}.

Kernels and quotients

  1. Show that \mathbb{Z}_{45}/\langle 9 \rangle \cong \mathbb{Z}_9, naming the homomorphism and its kernel.
  2. Show that f \mapsto (f(0), f(1)) is a ring homomorphism from \mathbb{Q}[x] onto \mathbb{Q} \times \mathbb{Q} with kernel \langle x^2 - x \rangle, and conclude \mathbb{Q}[x]/\langle x^2 - x \rangle \cong \mathbb{Q} \times \mathbb{Q}.
  3. Show that \mathbb{R}[x]/\langle x^2 - 2x + 2 \rangle \cong \mathbb{C} using f \mapsto f(1 + i), and find the representative a + bx of the coset of x^3 + x.
  4. Show that a + bi \mapsto [a + 5b]_{13} is a ring homomorphism from \mathbb{Z}[i] onto \mathbb{Z}_{13} with kernel \langle 3 + 2i \rangle.
  5. Show that f \mapsto [f(1)]_3 is a ring homomorphism from \mathbb{Z}[x] onto \mathbb{Z}_3 with kernel \langle 3, x - 1 \rangle. Do x^4 + x + 1 and x^5 + 2x + 7 lie in this ideal?

Equations with no solutions

  1. Show that x^2 - 5y^2 = 2 has no solution in integers.
  2. For the clue “a sum of the squares of three integers”, Amara proposes the entry 2{,}031. Show that it cannot be one.

Proofs

  1. Prove that the only ring homomorphism \mathbb{Q} \to \mathbb{R} is the inclusion, and that the identity is the only automorphism of \mathbb{Q}.
  2. Prove that there is no ring homomorphism \mathbb{Z}[i] \to \mathbb{R}, and none \mathbb{Z}_5 \to \mathbb{Z}.
  3. Prove that a ring homomorphism sends nilpotent elements to nilpotent elements. Using \mathbb{Z} \to \mathbb{Z}_4 and \mathbb{Z} \to \mathbb{Z}_5, show that an element that is not nilpotent can have a nonzero nilpotent image, and a non-unit can have a unit image.
  4. Let \psi : R \to S be onto with kernel K. Prove that J \mapsto \psi^{-1}(J) is a bijection from the ideals of S onto the ideals of R containing K, with inverse I \mapsto \psi(I). Use it to list the ideals of \mathbb{Z}_{12}.
  5. Let R be a commutative ring of prime characteristic p. Prove that the Frobenius map \operatorname{Fr}(a) = a^p is a ring homomorphism R \to R, one to one if R is an integral domain, and an automorphism if R is a finite integral domain.
  6. Prove that \displaystyle \mathbb{Z}[\sqrt{2}] has exactly two automorphisms.
Answers
  1. (a) No: 1 \mapsto -1. (b) Yes: (a + b)I_2 = aI_2 + bI_2, (ab)I_2 = (aI_2)(bI_2) and 1 \mapsto I_2; the kernel is \{0\}. (c) No: (1, 0)(0, 1) = (0, 0) \mapsto 0, but the images multiply to 1 \cdot 1 = 1. (d) No: \operatorname{tr} I_2 = 2 \ne 1.
  2. By Theorem 5, \mathbb{Z}_m \to \mathbb{Z}_n exists exactly when n \mid m. \mathbb{Z}_{12} \to \mathbb{Z}_4: [k]_{12} \mapsto [k]_4, kernel \{0, 4, 8\}. \mathbb{Z}_4 \to \mathbb{Z}_{12}: none, as 12 \nmid 4. \mathbb{Z}_{15} \to \mathbb{Z}_5: [k]_{15} \mapsto [k]_5, kernel \{0, 5, 10\}. \mathbb{Z} \to \mathbb{Z}_7: k \mapsto [k]_7, kernel 7\mathbb{Z}. \mathbb{Z}_7 \to \mathbb{Z}: none, as 7 \cdot 1 \ne 0 in \mathbb{Z}.
  3. k(1, 1) = ([k]_4, [k]_6) is 0 exactly when 4 \mid k and 6 \mid k, that is, 12 \mid k: characteristic 12, prime subring \{([k]_4, [k]_6)\}, with 12 of the 24 elements, \cong \mathbb{Z}_{12}. In \mathbb{Z}_5 \times \mathbb{Q}, k(1, 1) = ([k]_5, k) is 0 only for k = 0: characteristic 0, prime subring \{([k]_5, k)\} \cong \mathbb{Z}.
  4. Exactly two, (a, b) \mapsto a and (a, b) \mapsto b. With e = (1, 0), e^2 = e gives \psi(e)^2 = \psi(e) in \mathbb{Z}, so \psi(e) = 0 or 1, and \psi(0, 1) = \psi(1, 1) - \psi(e) = 1 - \psi(e). By Theorem 1(a), \psi(a, b) = a\psi(e) + b\psi(0, 1). Both projections work, as the operations are coordinatewise.
  5. As 9 \mid 45, [k]_{45} \mapsto [k]_9 is an onto ring homomorphism (Theorem 5) with kernel \{0, 9, 18, 27, 36\} = \langle 9 \rangle; apply Theorem 6.
  6. Evaluation respects sums and products (Example 4), the operations of \mathbb{Q} \times \mathbb{Q} are coordinatewise, and 1 \mapsto (1, 1). Onto: (a, b) is the image of a + (b - a)x. The kernel holds x^2 - x = x(x - 1); if f = q(x^2 - x) + c + dx has f(0) = c = 0 and f(1) = c + d = 0, then f \in \langle x^2 - x \rangle. Apply Theorem 6.
  7. (1 + i)^2 = 2i, so (1 + i)^2 - 2(1 + i) + 2 = 0. If f = q(x^2 - 2x + 2) + a + bx, then f(1 + i) = (a + b) + bi, which is 0 only for a = b = 0: the kernel is \langle x^2 - 2x + 2 \rangle. Onto: c + di is the image of (c - d) + dx. Then x^3 + x \mapsto (-2 + 2i) + (1 + i) = -1 + 3i, so b = 3, a = -4: the representative is -4 + 3x.
  8. Products: (ac - bd) + 5(ad + bc) and (a + 5b)(c + 5d) differ by 26bd, a multiple of 13; 1 \mapsto [1] and k \mapsto [k], so it is onto. 3 + 2i \mapsto [13] = 0. As (3 + 2i)(1 - i) = 5 - i and (3 + 2i)(3 - 2i) = 13, if 13 \mid a + 5b then a + bi = (a + 5b) - b(5 - i) \in \langle 3 + 2i \rangle.
  9. It is evaluation at 1 (Example 4) followed by reduction mod 3, a ring homomorphism by Theorem 1(d), onto via constants. It kills 3 and x - 1. Each x^k - 1 = (x - 1)(x^{k - 1} + \dots + 1), so f = (x - 1)g + f(1) with g \in \mathbb{Z}[x]; if 3 \mid f(1), then f \in \langle 3, x - 1 \rangle. At 1 the two polynomials give 3 and 10: the first lies in the ideal, the second does not.
  10. Reducing mod 5 (Theorem 1(b)) gives \psi(x)^2 = 2 in \mathbb{Z}_5, but the squares of 0, 1, 2, 3, 4 are 0, 1, 4, 4, 1.
  11. Squares mod 8 are 0, 1, 4, so sums of three are 0, 1, 2, 3, 4, 5, 6 (for instance 4 + 4 + 4 = 12 \equiv 4), never 7. But 2031 = 8 \cdot 253 + 7.
  12. Proof. Let \psi : \mathbb{Q} \to \mathbb{R} be a ring homomorphism. By Theorem 5(a), \psi(m) = m for integers m. For n \ge 1, \displaystyle n \cdot \frac{m}{n} = m, so \displaystyle n\psi(\frac{m}{n}) = \psi(m) = m (Theorem 1(a)) and \displaystyle \psi(\frac{m}{n}) = \frac{m}{n}. The same argument with \mathbb{Q} as the target shows an automorphism of \mathbb{Q} is the identity. \square
  13. Proof. If \psi : \mathbb{Z}[i] \to \mathbb{R}, then \psi(i)^2 = \psi(i^2) = \psi(-1) = -1, and no real square is negative. A map \mathbb{Z}_5 \to \mathbb{Z} would need 5 \cdot 1 = 0 in \mathbb{Z} (Theorem 5(b)). \square
  14. Proof. If a^k = 0, then \psi(a)^k = \psi(a^k) = \psi(0) = 0 (Theorem 1(a)). \square 2 \in \mathbb{Z} is not nilpotent, but [2]_4 \ne 0 has [2]_4^2 = 0. 2 is not a unit of \mathbb{Z}, but [2]_5[3]_5 = [1]_5.
  15. Proof. \psi^{-1}(J) is an ideal containing K (Theorem 2(b)), and \psi(\psi^{-1}(J)) = J as \psi is onto. For an ideal I \supseteq K, \psi(I) is an ideal (Theorem 2(c)), and \psi^{-1}(\psi(I)) = I: if \psi(a) = \psi(b) with b \in I, then a - b \in K \subseteq I, so a \in I. The two maps are inverse bijections. \square With \pi : \mathbb{Z} \to \mathbb{Z}_{12}, the ideals of \mathbb{Z} containing 12\mathbb{Z} are d\mathbb{Z} with d \mid 12, so \mathbb{Z}_{12} has six ideals \langle d \rangle, d = 1, 2, 3, 4, 6, 12.
  16. Proof. (ab)^p = a^pb^p by commutativity, (a + b)^p = a^p + b^p (Rings and integral domains, Exercise 15), and 1^p = 1. In an integral domain a^p = 0 forces a = 0, as a product of nonzero elements is nonzero; so the kernel is \{0\} and \operatorname{Fr} is one to one (Theorem 3). A one-to-one map of a finite set to itself is onto (Result 3 of One-to-one, onto and bijective functions). \square
  17. Proof. An automorphism \psi fixes integers (Theorem 5(a)). Write \displaystyle \psi(\sqrt{2}) = u + v\sqrt{2}; then \displaystyle (u^2 + 2v^2) + 2uv\sqrt{2} = \psi(2) = 2, so uv = 0 and u^2 + 2v^2 = 2 (as in Example 6), forcing u = 0, v = \pm 1. So \displaystyle \psi(a + b\sqrt{2}) = a \pm b\sqrt{2}. The identity is an automorphism, and so is \displaystyle \sigma(a + b\sqrt{2}) = a - b\sqrt{2}: it respects sums, sends 1 to 1, is its own inverse, and \displaystyle (a + b\sqrt{2})(c + d\sqrt{2}), which is \displaystyle (ac + 2bd) + (ad + bc)\sqrt{2}, maps to \displaystyle (ac + 2bd) - (ad + bc)\sqrt{2}, which is \displaystyle (a - b\sqrt{2})(c - d\sqrt{2}). \square
math.wiki · updated 2026-11-17 · CC-BY-SA-4.0