fact that every prime ideal is a radical ideal (Section 8.3, Problem 2).] No other

decomposition is possible, for if V (fi ) ⊆ V (fj ), then (fi ) ⊇ (fj ). This is impossible

if fi and fj are distinct irreducible factors of f .

19

2. By the Nullstellensatz, IV (f ) = (f ), and we claim that (f ) = (f1 · · · fr ). For if

g ∈ (f1 · · · fr ), then a su¬ciently high power of g will belong to (f). Conversely, if

g m = hf , then each fi divides g m , and since the fi are irreducible, each fi divides g,

so (f1 · · · fr ) divides g.

3. By Problem 1, f is irreducible if and only if V (f ) is an irreducible hypersurface. If f

and g are irreducible and V (f ) = V (g), then as in Problem 1, (f ) = (g), so f = cg

for some nonzero constant c (Section 2.1, Problem 2). Thus f ’ V (f ) is a bijection

between irreducible polynomials and irreducible hypersurfaces, if the polynomials f

and cf, c = 0, are identi¬ed.

4. This follows from the de¬nition of I(X) in (8.1.3), and the observation that a function

vanishes on a union of sets i¬ it vanishes on each of the sets.

5. By Section 8.1, Problem 5, every variety V is the union of ¬nitely many irreducible

subvarieties V1 , . . . , Vr . By Problem 4, I(V ) = ©r I(Vi ). By the Problems in Sec-

i=1

tion 8.1, each I(Vi ) is a prime ideal. By (8.4.3), every radical ideal is I(V ) for some

variety V , and the result follows.

6. By Section 8.1, Problem 6, and the inclusion-reversing property of I (part (4) of

(8.1.3)), the decomposition is unique if we discard any prime ideal that properly

contains another one. In other words, we retain only the minimal prime ideals.

7. If f is any irreducible factor of any of the fi , then f does not divide g. Thus for some

j = i, f does not divide fj . By Problem 7 of Section 8.3, the simultaneous equations

f = fj = 0 have only ¬nitely many solutions, and consequently X is a ¬nite set.

8. With notation as in Problem 7, fi = ghi , where the gcd of the hi is constant. Thus

X is the union of the algebraic curve de¬ned by g = 0 and the ¬nite set de¬ned by

h1 = · · · = hm = 0. (This analysis does not apply when X is de¬ned by the zero

polynomial, in which case X = A2 .)

9. If k = R, the zero-set of x2 + y 2n is {(0, 0)} for all n = 1, 2, . . . . If k is algebraically

closed, then as a consequence of the Nullstellensatz, V (f ) = V (g) with f and g

irreducible implies that f = cg for some constant c. (See Problem 3).

10. Let k = F2 , and let I be the ideal of k[X] generated by f (X) = X 2 + X + 1. Since

f is irreducible, I is a maximal ideal (Section 3.1, Problem 8), in particular, I is

proper. But f (0) and f (1) are nonzero, so V (I) is empty, contradicting the weak

Nullstellensatz.

Section 8.5

1. If x ∈ M , then the ideal generated by M and x is R, by maximality of M . Thus there

/

exists y ∈ M and z ∈ R such that y + zx = 1. By hypothesis, zx, hence x, is a unit.

Take the contrapositive to conclude that every nonunit belongs to M .

2. Any additive subgroup of the cyclic additive group of Zpn must consist of multiples

of some power of p, and it follows that every ideal is contained in (p), which must

therefore be the unique maximal ideal.

3. No. A can be nilpotent, that is, some power of A can be 0. The set will be multiplicative

if A is invertible.

20

4. S ’1 (gf ) takes m/s to g(f (m))/s, as does S ’1 gS ’1 f . If f is the identity on M , then

S ’1 f is the identity on S ’1 M .

5. By hypothesis, gf = 0, so S ’1 gS ’1 f = S ’1 gf = S ’1 0 = 0. Thus im S ’1 f ⊆ ker S ’1 g.

Conversely, let x ∈ N, s ∈ S, with x/s ∈ ker S ’1 g. Then g(x)/s = 0/1, so for some

t ∈ S we have tg(x) = g(tx) = 0. Therefore tx ∈ ker g = im f , so tx = f (y) for some

y ∈ M . We now have x/s = f (y)/st = (S ’1 f )(y/st) ∈ im S ’1 f .

6. The set of nonuits is M = {f /g : g(a) = 0, f (a) = 0}, which is an ideal. By (8.5.9), R

is a local ring with maximal ideal M .

7. The sequence 0 ’ N ’ M ’ M/N ’ 0 is exact, so by Problem 5, 0 ’ NS ’

MS ’ (M/N )S ’ 0 is exact. (If f is one of the maps in the ¬rst sequence, the

corresponding map in the second sequence is S ’1 f .) It follows from the de¬nition of

localization of a module that NS ¤ MS , and by exactness of the second sequence we

have (M/N )S ∼ MS /NS , as desired.

=

Section 8.6

√ √

1. If xm belongs to the intersection of the Ii , then x belongs to each I i , so x ∈ ©n

i=1 I i .

√

Conversely, if x ∈ ©i=1 I i , let x ∈ Ii . If m is the maximum of the mi , then

n mi

x ∈ ©i=1 Ii , so x ∈ ©i=1 i

m n n I.

2. We are essentially setting X = Z = 0 in R, and this collapses R down to k[Y ]. Formally,

map f + I to g + I, where g consists of those terms in f that do not involve X or Z.

Then R/P ∼ k[Y ], an integral domain. Therefore P is prime.

=

√

3. (X + I)(Y + I) = Z 2 + I ∈ P 2 , but X + I ∈ P 2 and Y + I ∈ P 2 = P .

/ /

4. P1 is prime because R/P1 ∼ k[Y ], an integral domain. P2 is maximal by (8.3.1), so P2 2

=

is P2 -primary by (8.6.6). The radical of Q is P2 , so by (8.6.5), Q is P2 -primary.

5. The ¬rst assertion is that

(X 2 , XY ) = (X) © (X, Y )2 = (X) © (X 2 , XY, Y 2 )

and the second is

(X 2 , XY ) = (X) © (X 2 , Y ).

In each case, the left side is contained in the right side by de¬nition of the ideals

involved. The inclusion from right to left follows because if f (X, Y )X = g(X, Y )Y 2

(or f (X, Y )X = g(X, Y )Y ), then g(X, Y ) must involve X and f (X, Y ) must involve Y .

Thus f (X, Y )X is a polynomial multiple of XY .

6. By (8.6.9), a proper ideal I can be expressed as the intersection of ¬nitely many primary

ideals Qi . If Qi is Pi -primary, then by Problem 1,

√

I = I = ©i Qi = ©i Pi .

√ √

have X, Y ∈ In , so (X, Y ) ⊆ In . By (8.3.1),

7. Since X 3 and Y n belong to In , we √

(X, Y ) is √ maximal ideal. Since In is proper (it does not contain 1), we have

a

(X, Y ) = In . By (8.6.5), In is primary.

21

Section 8.7

1. [(x, y + y ) + G] ’ [(x, y) + G] ’ [(x, y ) + G] = 0 since (x, y + y ) ’ (x, y) ’ (x, y ) ∈ G;

r[(x, y) + G] ’ [(rx, y) + G] = 0 since r(x, y) ’ (rx, y) ∈ G; the other cases are similar.

2. Let a and b be integers such that am + bn = 1. If x ∈ Zm and y ∈ Zn , then

x — y = 1(x — y) = a(mx — y) + b(x — ny) = 0 since z — 0 = 0 — z = 0.

3. Let A be a torsion abelian group, that is, each element of A has ¬nite order. If x ∈ A

and y ∈ Q, then nx = 0 for some positive integer n. Thus x — y = n(x — (y/n)) =

nx — (y/n) = 0 — (y/n) = 0.

4. We have h = g h = g gh and h = gh = gg h . But if P = T and f = h, then

g = 1T makes the diagram commute, as does g g. By the uniqueness requirement in

the universal mapping property, we must have g g = 1T , and similarly gg = 1T . Thus

T and T are isomorphic.

5. n — x = n(1 — x) = 1 — nx = 1 — 0 = 0.

6. nZ — Zn ∼ Z — Zn ∼ Zn by (8.7.6), with n — x ’ 1 — x ’ x, and since x = 0, n — x

= =

cannot be 0.

7. We have a bilinear map (f, g) ’ f — g from HomR (M, M ) — HomR (N, N ) to

HomR (M —R N, M —R N ), and the result follows from the universal mapping property

of tensor products.

8. In terms of matrices, we are to prove that Mm (R) — Mn (R) ∼ Mmn (R). This follows

=

2

because Mt (R) is a free R-module of rank t .

Section 8.8

1. y1 · · · (yi +yj ) · · · (yi +yj ) · · · yp = y1 · · · yi · · · yi · · · yp +y1 · · · yi · · · yj · · · yp +y1 · · · yj · · ·

yi · · · yp + y1 · · · yj · · · yj · · · yp . The left side, as well as the ¬rst and last terms on the

right, are zero by de¬nition of N . Thus y1 · · · yi · · · yj · · · yp = ’y1 · · · yj · · · yi · · · yp , as

asserted.

2. If π is any permutation of {a, . . . , b}, then by Problem 1,

xπ(a) · · · xπ(b) = (sgn π)xa · · · xb .

The left side will be ±xa · · · xb , regardless of the particular permutation π, and the

result follows.

3. The multilinear map f induces a unique h : M —p ’ Q such that h(y1 — · · · — yp ) =

f (y1 , . . . , yp ). Since f is alternating, the kernel of h contains N , so the existence and

uniqueness of the map g follows from the factor theorem.

4. By Problem 3, there is a unique R-homomorphism g : Λn M ’ R such that

g(y1 · · · yn ) = f (y1 , . . . , yn ). In particular, g(x1 · · · xn ) = f (x1 , . . . , xn ) = 1 = 0. Thus

x1 · · · xn = 0. If r is any nonzero element of R, then g(rx1 x2 · · · xn ) = f (rx1 , . . . , xn ) = r,

so rx1 · · · xn = 0. By Problem 2, {x1 · · · xn } is a basis for Λn M .

I aI xI = 0, xI ∈ R,

5. Fix the set of indices I0 and its complementary set J0 . If

multiply both sides on the right by xJ0 . If I = I0 , then xI xJ0 = 0 by de¬nition of N .

22

Thus aI0 xI0 xJ0 = ±aI0 x1 · · · xn = 0. By Problem 4, aI0 = 0. Since I0 is arbitrary, the

result follows.

6. We have R0 ⊆ S by de¬nition of S. Assume that Rm ⊆ S for m = 0, 1, . . . , n ’ 1, and

r

let a ∈ Rn (n > 0). Then a ∈ I, so a = i=1 ci xi where (since xi ∈ Rni and R is the

direct sum of the Rm ) ci ∈ Rn’ni . By induction hypothesis, ci ∈ S, and since xi ∈ S

by de¬nition of S, we have a ∈ S, completing the induction.

7. The “if” part follows from Section 8.2, Problem 5, so assume R Noetherian. Since

R0 ∼ R/I, it follows that R0 is Noetherian. Since R is Noetherian, I is ¬nitely

=

generated, so by Problem 6, R = S, a ¬nitely generated R0 -algebra.

Section 9.1

1. Assume R is simple, and let x ∈ R, x = 0. Then Rx coincides with R, so 1 ∈ Rx. Thus

there is an element y ∈ R such that yx = 1. Similarly, there is an element z ∈ R such

that zy = 1. Therefore

z = z1 = zyx = 1x = x, so xy = zy = 1

and y is a two-sided inverse of x. Conversely, assume that R is a division ring, and

let x be a nonzero element of the left ideal I. If y is the inverse of x, then 1 = yx ∈ I,

so I = R and R is simple.

I is proper because f (1) = x = 0, and R/I ∼ Rx by the ¬rst isomorphism theorem.

2. =

3. The “if” part follows from the correspondence theorem, so assume that M is simple.

If x is a nonzero element of M , then M = Rx by simplicity. If I = ker f as in

Problem 2, then M ∼ R/I, and I is maximal by the correspondence theorem.

=

4. The “only if” part was done in Problem 3, so assume that M is not simple. Let N be a

submodule of M with 0 < N < M . If x is a nonzero element of N , then Rx ¤ N < M ,

so x cannot generate M .

By Problem 3, a simple Z-module is isomorphic to Z/I, where I is a maximal ideal

5.

of Z. By Section 2.4, Problems 1 and 2, I = (p) where p is prime.

6. As in Problem 5, a simple F [X]-module is isomorphic to F [X]/(f ), where f is an

irreducible polynomial in F [X]. (See Section 3.1, Problem 8.)

7. If x is a nonzero element of V and y an arbitrary element of V , there is an endomor-

phism f such that f (x) = y. Therefore V = (Endk V )x. By Problem 4, V is a simple

Endk (V )-module.

By (4.7.4), every such short exact sequence splits i¬ for any submodule N ¤ M ,

8.

M ∼ N • P , where the map N ’ M can be identi¬ed with inclusion and the map

=

M ’ P can be identi¬ed with projection. In other words, every submodule of M is a

direct summand. Equivalently, by (9.1.2), M is semisimple.

Section 9.2

1. Unfortunately, multiplication by r is not necessarily an R-endomorphism of M , since

r(sx) = (rs)x, which need not equal s(rx) = (sr)x.

23

2. Let x be a generator of M , and de¬ne f : R ’ M by f (r) = rx. By the ¬rst isomor-

phism theorem, M ∼ R/ ann M . The result follows from the correspondence theorem.

=

3. Let M = Zp •Zp where p is prime. Then M is not a simple Z-module, but ann M = pZ

is a maximal ideal of Z.

4. The computation given in the statement of the problem shows that (1,0) is a generator

of V , hence V is cyclic. But N = {(0, b) : b ∈ F } is a nontrivial proper submodule of

V . (Note that T (0, b) = (0, 0) ∈ N .) Therefore V is not simple.

5. Since F is algebraically closed, f has an eigenvalue » ∈ F . Thus the kernel of f ’ »I

is not zero, so it must be all of M . Therefore f (m) = »m for all m ∈ M .

6. If r ∈ I and s + I ∈ R/I, then r(s + I) = rs + I. But if I is not a right ideal, we cannot

guarantee that rs belongs to I.

Section 9.3

1. For each j = 1, . . . , n, there is a ¬nite subset I(j) of I such that xj belongs to the

direct sum of the Mi , i ∈ I(j). If J is the union of the I(j), then M ⊆ i∈J Mi ⊆ M ,

so M is the direct sum of the Mi , i ∈ J.

2. Each simple module Mi , i = 1, . . . , n, is cyclic (Section 9.1, Problem 4), and therefore

can be generated by a single element xi . Thus M is generated by x1 , . . . , xn .

3. A left ideal is simple i¬ it is minimal, so the result follows from (9.1.2).

4. No. If it were, then by Section 9.1, Problem 5, Z would be a direct sum of cyclic groups

of prime order. Thus each element of Z would have ¬nite order, a contradiction.

5. By (4.6.4), every ¬nite abelian group is the direct sum of various Zp , p prime. If p and

q are distinct primes, then Zp • Zq ∼ Zpq by the Chinese remainder theorem. Thus Zn

=

can be assembled from cyclic groups of prime order as long as no prime appears more

than once in the factorization of n. (If Zp • Zp is part of the decomposition, the group

cannot be cyclic.) Consequently, Zn is semisimple if and only if n is square-free.

6. This follows from Section 9.1, Problem 8. (In the ¬rst case, B is semisimple by hy-

pothesis, and in the second case A is semisimple. The degenerate case M = 0 can be

handled directly.)

7. Conditions (a) and (b) are equivalent by (9.3.2) and the de¬nition of semisimple ring.

By Problem 6, (b) implies both (c) and (d). To show that (c) implies (b) and (d)

implies (b), let M be a nonzero R-module, with N a submodule of M . By hypothesis,

the sequence 0 ’ N ’ M ’ M/N ’ 0 splits. (By hypothesis, M/N is projective in

the ¬rst case and N is injective in the second case.) By Section 9.1, Problem 8, M is

semisimple.

Section 9.4

1. If there is an in¬nite descending sequence I1 ⊃ I2 ⊃ · · · of left ideals, we can proceed

exactly as in (9.4.7) to reach a contradiction.

24

2. Let I1 be any nonzero left ideal. If I1 is simple, we are ¬nished. If not, there is

a nonzero left ideal I2 such that I1 ⊃ I2 . If we continue inductively, the Artinian

hypothesis implies that the process must terminate in a simple left ideal It .

3. By Problem 2, the ring R has a simple R-module M . The hypothesis that R has no

nontrivial two-sided ideals implies that we can proceed exactly as in (9.4.6) to show

that M is faithful.

4. If V is in¬nite-dimensional over D, then exactly as in (9.4.7), we ¬nd an in¬nite

descending chain of left ideals, contradicting the assumption that R is Artinian.

5. By Problem 4, V is a ¬nite-dimensional vector space over D, so we can reproduce the

discussion preceding (9.4.7) to show that R ∼ EndD (V ) ∼ Mn (Do ).

= =

6. The following is a composition series:

0 < M1 < M1 • M2 < · · · < M1 • M2 • · · · • Mn = M.

7. By (9.1.2), M is a direct sum of simple modules. If the direct sum is in¬nite, then we

can proceed as in Problem 6 to construct an in¬nite ascending (or descending) chain

of submodules of M , contradicting the hypothesis that M is Artinian and Noetherian.

Section 9.5

1. If g ∈ ker ρ, then gv = v for every v ∈ V . Take v = 1G to get g1G = 1G , so g = 1G

and ρ is injective.

2. (gh)(v(i)) = v(g(h(i))) and g(h(v(i))) = g(v(h(i))) = v(g(h(i))). Also, 1G (v(i)) =

v(1G (i)) = v(i).

3. We have g(v(1)) = v(4), g(v(2)) = v(2), g(v(3)) = v(1), g(v(4)) = v(3), so

®

0010

0 1 0 0

[g] =

°0 0 0 1» .

1000

4. We have gv1 = v2 , gv2 = v3 = ’v1 ’ v2 ; hv1 = v1 , hv2 = v3 = ’v1 ’ v2 . Thus

’1 1 ’1

0

[g] = , h= .

’1 0 ’1

1

√

5. We have v1 = e1 and v2 = ’ 1 e1 + 1 3e2 . Thus

2 2

√

® ®

1 ’1 113

2 3

P ’1 = ° », P = ° »

√ √

013 023

2 3

and

√

®

’1 ’1 3

2 2

[g] = P ’1 [g]P = ° √ »,

’1

1

3

2 2

25

1 0

[h] = P ’1 [h]P = .

’1

0

6. Check by direct computation that the matrices A and B satisfy the de¬ning relations

of D8 : A4 = I, B 2 = I, AB = BA’1 . (See Section 5.8.)

7. Yes. Again, check by direct computation that the matrices Ai B j , i = 0, 1, 2, 3, j = 0, 1,

are distinct. Thus if g ∈ D8 and ρ(g) = I, then g is the identity element of D8 .

Section 9.6

1. Let W be the one-dimensional subspace spanned by v1 +v2 +v3 . Since any permutation

in S3 permutes the vi , v ∈ W implies gv ∈ W .

1 0

2. Multiplying [ar ] by , we have ar v1 = v1 and ar v2 = rv1 + v2 . Since W is

and

0 1

spanned by v1 , it is closed under the action of G and is therefore a kG-submodule.

3. If

x x

[ar ] =c

y y

then x + ry = cx and y = cy. If y = 0 then c = 1, so ry, hence y, must be 0. Thus

c = 1 and x is arbitrary, so that any one-dimensional kG-submodule must coincide

with W .

4. If V = W • U , where U is a kG-submodule, then U must be one-dimensional. By

Problem 3, W = U , and since W = 0, this is impossible.

5. If M is semisimple and either Noetherian or Artinian, then M is the direct sum of

¬nitely many simple modules. These simple modules are the factors of a composition

series, and the result follows from (7.5.12).

6. Let e be the natural projection on M1 . Then e is an idempotent, e = 0 since M1 = 0,

and e = 1 since e = 0 on M2 and M2 = 0.

7. Let e be a nontrivial idempotent, and de¬ne e1 = e, e2 = 1 ’ e. By direct computation,

e1 and e2 are nontrivial idempotents that are orthogonal. Take M1 = e1 (M ), M2 =

e2 (M ). Then M1 and M2 are nonzero submodules with M = M1 + M2 . To show that

the sum is direct, let z = e1 x = e2 y ∈ M1 © M2 , with x, y ∈ M . Then e1 z = e1 e2 y = 0,

and similarly e2 z = 0. Thus z = 1z = e1 z + e2 z = 0.

Section 9.7

1. If M = Rx, de¬ne f : R ’ M by f (r) = rx. By the ¬rst isomorphism theorem,

M ∼ R/ ker f . Moreover, ker f = ann(M ). Conversely, R/I is cyclic since it is

=

generated by 1 + I.

2. If N is a maximal submodule of M , then N is the kernel of the canonical map of M

onto the simple module M/N . Conversely, if f : M ’ S, S simple, then f (M ) is 0

26

or S, so f = 0 or S ∼ M/ ker f . Thus ker f is either M or a maximal submodule of M .

=

The intersection of all kernels therefore coincides with the intersection of all maximal

submodules. [If there are no maximal submodules, then the intersection of all kernels

is M .]

3. By the correspondence theorem, the intersection of all maximal left ideals of R/I is 0.

This follows because the intersection of all maximal left ideals of R containing I is

the intersection of all maximal left ideals of R. [Note that J(R) is contained in every

maximal left ideal, and I = J(R).]

4. Let g be an R-module homomorphism from N to the simple R-module S. Then

gf : M ’ S, so by Problem 2, J(M ) ⊆ ker(gf ). But then f (J(M )) ⊆ ker g. Take the

intersection over all g to get f (J(M )) ⊆ J(N ).

Suppose a ∈ J(R). If 1+ab is a nonunit, then it belongs to some maximal ideal M . But

5.

then a belongs to M as well, and therefore so does ab. Thus 1 ∈ M , a contradiction.

Now assume a ∈ J(R), so that for some maximal ideal M , a ∈ M . By maximality,

/ /

M + Ra = R, so 1 = x + ra for some x ∈ M and r ∈ R. Since x belongs to M , it

cannot be a unit, so if we set b = ’r, it follows that 1 + ab is a nonunit.

6. By the correspondence theorem, there is a bijection, given by ψ(A) = A/N , between

maximal submodules of M containing N and maximal submodules of M/N . Since

N ¤ J(M ) by hypothesis, a maximal submodule of M containing N is the same

thing as a maximal submodule of M . Thus J(M ) corresponds to J(M/N ), that is,

ψ(J(M )) = J(M/N ). Since ψ(J(M )) = J(M )/N , the result follows.

Section 9.8

1. at ∈ (at+1 ), so there exists b ∈ R such that at = bat+1 . Since R is an integral domain

we have 1 = ba.

2. Let a be a nonzero element of the Artinian integral domain R. The sequence (a) ⊇

(a2 ) ⊇ . . . stabilizes, so for some t we have (at ) = (at+1 ). By Problem 1, a has an

inverse, proving that R is a ¬eld.

3. If P is a prime ideal of R, then R/P is an Artinian integral domain by (7.5.7) and

(2.4.5). By Problem 2, R/P is a ¬eld, so by (2.4.3), P is maximal.

4. We have P © I ∈ S and P © I ⊆ I, so by minimality of I, P © I = I. Thus

P ⊇ I = ©n Ij , so by (7.6.2), P ⊇ Ij for some j. But P and Ij are maximal ideals,

j=1

hence P = Ij .

5. If z ∈ F , with z = x + y, x ∈ M , y ∈ N , de¬ne h : F ’ M/JM • N/JN by

h(z) = (x + JM ) + (y + JN ). Then h is an epimorphism with kernel JM + JN = JF ,

and the result follows from the ¬rst isomorphism theorem.

6. By (9.8.5), M/JM is an n-dimensional vector space over the residue ¬eld k. Since F ,

hence F/JF , is generated by n elements, F/JF has dimension at most n over k.

Thus N/JN must have dimension zero, and the result follows.

7. Multiplication of an element of I on the right by a polynomial f (X, Y ) amounts to

multiplication by the constant term of f . Thus I is ¬nitely generated as a right R-

module i¬ it is ¬nitely generated as an abelian group. This is a contradiction, because

27

as an abelian group,

I = •∞ ZX n Y.

n=0

8. By the Hilbert basis theorem, Z[X] is a Noetherian ring and therefore a Noetherian

left Z[X]-module. The isomorphic copy Z[X]Y is also a Noetherian left Z[X]-module,

hence so is R, by (7.5.8). A left ideal of R that is ¬nitely generated as a Z[X]-module

is ¬nitely generated as an R-module, so R is left-Noetherian.

9. The set {x1 , . . . , xn } spans V , and therefore contains a basis. If the containment is

proper, then by (9.8.5) part (ii), {x1 , . . . , xn } cannot be a minimal generating set, a

contradiction.

10. In vector-matrix notation we have y = Ax and therefore y = Ax, where aij = aij + J.

By Problem 9, x and y are bases, so that det A = 0. But under the canonical map of

R onto k, det A maps to det A, and therefore det A cannot belong to the kernel of

the map, namely J. But by (8.5.9), J is the ideal of nonunits of R, so det A is a unit.

Section 10.1

1. This is exactly the same as the proof for groups in (1.1.1).

2. Any ring homomorphism on Q is determined by its values on Z (write m = (m/n)n

and apply the homomorphism g to get g(m/n) = g(m)/g(n). Thus if gi = hi, then g

coincides with h on Z, so g = h, proving that i is epic.

3. Let AZB be a shorthand notation for the composition of the unique morphism from

A to the zero object Z, followed by the unique morphism from Z to B. If Z is another

zero object, then AZB = (AZ Z)B = A(Z ZB) = A(Z B) = AZ B, as claimed.

4. If is = it, then f is = f it = 0, so is(= it) = ih where h is unique. Thus s and t must

coincide with h, hence i is monic.

5. A kernel of a monic f : A ’ B is 0, realized as the zero map from a zero object Z

to A. For f 0 = 0, and if f g = 0, then f g = f 0; since f is monic, g = 0. But then g

can be factored through 0ZA . Similarly, a cokernel of the epic f : A ’ B is the zero

map from B to a zero object.

6. If j = ih and i = jh , then i = ihh , so by uniqueness in part (2) of the de¬nition of

kernel (applied when g = i), hh , and similarly h h, must be the identity.

7. De¬ne h : K ’ A by h(x) = 1 for all x. Since K is nontrivial, h cannot be injective,

so that g = h. But f g = f h, since both maps take everything in K to the identity

of B.

8. The kernel of a ring homomorphism is an ideal, but not a subring (since it does not

contains the multiplicative identity).

9. Let f : A ’ B be a noninjective ring homomorphism. Let C be the set of pairs (x, y)

in the direct product A — A such that f (x) = f (y). Since f is not injective, there is

an element (x, y) of C with x = y. Thus if D = {(x, x) : x ∈ A}, then D ‚ C. If g is

the projection of A — A on the ¬rst coordinate, and h is the projection on the second

coordinate, then f (g(x, y)) = f (x) and f (h(x, y)) = f (y), so f g = f h on the ring C.

But g and h disagree on the nonempty set C \ D, so f is not monic.

28

10. Let g be the canonical map of N onto N/f (M ), and let h : N ’ N/f (M ) be identically

zero. Since gf sends everything to 0, we have gf = hf with g = h. Thus f is not epic.

Section 10.2

1. Suppose that y is the product of the xi . By de¬nition of product, if fi : x ’ xi for

all i, there is a unique f : x ’ y such that pi f = fi . Since pi : y ’ xi , we have y ¤ xi .

Moreover, if x ¤ xi for all i, then x ¤ y. Therefore y is a greatest lower bound of

the xi .

2. No. For example, consider the usual ordering on the integers.

3. By duality, a coproduct of the xi , if it exists, is a least upper bound.

4. If x has order r and y has order s, then rs(x + y) = s(rx) + r(sy) = 0. Thus the sum

of two elements of ¬nite order also has ¬nite order, and the result follows.

5. The key point is that if f is a homomorphism of a torsion abelian group S, then f (S)

is a also torsion [since nf (x) = f (nx)]. Thus in diagram (1) with A = Ai , we have

f (S) ⊆ T (A). Since Ai is the product in the category of abelian groups, it follows

that T (A) satis¬es the universal mapping property and is therefore the product in the

category of torsion abelian groups.

6. Given homomorphisms fj : Gj ’ H, we must lift the fj to a homomorphism from

the free product to H. This is done via f (a1 · · · an ) = f1 (a1 ) · · · fn (an ). If ij is the

inclusion map from Gj to —i Gi , then f (ij (aj )) = f (aj ) = fj (aj ), as required.

7. We have pi f = fi , where fi : G ’ Ci . The fi can be chosen to be surjective (e.g., take

G to be the direct product of the Ci ), and it follows that the pi are surjective.

8. Since f : C1 ’ C, we have f (a1 ) = na for some positive integer n. Thus

a1 = f1 (a1 ) = p1 f (a1 ) = p1 (na) = np1 (a) = na1 ;

0 = f2 (a1 ) = p2 f (a1 ) = p2 (na) = np2 (a) = na2 .

9. By Problem 8, the order of C1 divides n ’ 1, and the order of C2 divides n. There are

many choices of C1 and C2 for which this is impossible. For example, let C1 and C2

be nontrivial p-groups for a ¬xed prime p.

Section 10.3

1. If f : x ’ y, then F f : F x ’ F y. By de¬nition of the category of preordered sets, this

statement is equivalent to x ¤ y =’ F x ¤ F y. Thus functors are order-preserving

maps.

2. F must take the morphism associated with xy to the composition of the morphism

associated with F x and the morphism associated with F y. In other words, F (xy) =

F (x)F (y), that is, F is a homomorphism.

3. If β ∈ X — , then (gf )— (β) = βgf and f — g — (β) = f — (βg) = βgf .

29

4. To verify the functorial property, note that

(gf )—— (v —— ) = v —— (gf )— = v —— f — g — (by Problem 3)

and

g —— f —— v —— = g —— (v —— f — ) = v —— f — g — .

Thus (gf )—— = g —— f —— . If f is the identity, then so is f — , and consequently so is f —— .

5. f —— tV (v) = f —— (v) = vf — , and if β ∈ W — , then (vf — )(β) = v(f — β) = (f — β)(v) = βf (v).

But tW f (v) = f (v) where f (v)(β) = βf (v).

6. Groups form a subcategory because every group is a monoid and every group homo-

morphism is, in particular, a monoid homomorphism. The subcategory is full because

every monoid homomorphism from one group to another is also a group homomor-

phism.

7. (a) If two group homomorphisms are the same as set mappings, they are identical as

homomorphisms as well. Thus the forgetful functor is faithful. But not every map

of sets is a homomorphism, so the forgetful functor is not full.

(b) Since (f, g) is mapped to f for arbitrary g, the projection functor is full but not

faithful (except in some degenerate cases).

Section 10.4

1. If a homomorphism from Z2 to Q takes 1 to x, then 0 = 1 + 1 ’ x + x = 2x. But 0

must be mapped to 0, so x = 0.

2. A nonzero homomorphism can be constructed with 0 ’ 0, 1 ’ Then 1 + 1 ’

1

2.

2 + 2 = 1 = 0 in Q/Z.

1 1

3. Since a trivial group cannot be mapped onto a nontrivial group, there is no way that

F g can be surjective.

4. Let f be a homomorphism from Q to Z. If r is any rational number and m is a positive

integer, then

r r r

+ ··· +

f (r) = f = mf

m m m

so

r f (r)

f = .

m m

But if f (r) = 0, we can choose m such that f (r)/m is not an integer, a contradiction.

Therefore f = 0.

5. By Problem 4, Hom(Q, Z) = 0. But Hom(Z, Z) = 0, so as in Problem 3, Gf cannot be

surjective.

30

6. We have Z2 — Z ∼ Z2 and Z2 — Q = 0

=

m 2m m m

1— =1— =2— =0— = 0.

n 2n 2n 2n

Thus the map Hf cannot be injective.

7. Since f— = F f is injective, f ± = 0 implies ± = 0, so f is monic and hence injective.

Since g— f— = 0, we have gf ± = 0 for all ± ∈ Hom(M, A). Take M = A and ± = 1A to

conclude that gf = 0, so that im f ⊆ ker g. Finally, take M = ker g and ± : M ’ B

the inclusion map. Then g— ± = g± = 0, so ± ∈ ker g— = im f— . Thus ± = f β for some

β ∈ Hom(M, A). Thus ker g = M = im ± ⊆ im f .

8. If (3) is exact for all possible R-modules N , then (1) is exact. This is dual to the result

of Problem 7, and the proof amounts to interchanging injective and surjective, monic

and epic, inclusion map and canonical map, kernel and cokernel.

Section 10.5

1. If x ∈ P , then f (x) can be expressed as i ti ei (a ¬nite sum), and we de¬ne fi (x) = ti

and xi = π(ei ). Then

x = π(f (x)) = π( ti ei ) = ti π(ei ) = fi (x)xi .

i i i

2. π(f (x)) = fi (x)π(ei ) = fi (x)xi = x.

i i

3. By Problem 2, the exact sequence 0 ’ ker π ’ F ’ P ’ 0 (with π : F ’ P ) splits,

and therefore P is a direct summand of the free module F and hence projective.

4. Since Rn is free, the “if” part follows from (10.5.3), part (4). If P is projective, then

by the proof of (3) implies (4) in (10.5.3), P is a direct summand of a free module of

rank n. [The free module can be taken to have a basis whose size is the same as that

of a set of generators for P .]

5. This follows from Problem 1 with F = Rn .

6. If P is projective and isomorphic to M/N , we have an exact sequence 0 ’ N ’

M ’ P ’ 0. Since P is projective, the sequence splits by (10.5.3), part (3), so P is

isomorphic to a direct summand of M . Conversely, assume that P is a direct summand

of every module of which it is a quotient. Since P is a quotient of a free module F , it

follows that P is a direct summand of F . By (10.5.3) part (4), P is projective.

7. In the diagram above (10.5.1), take M = R1 • R2 , with R1 = R2 = R. Take P =

N = R1 , f = 1R1 , and let g be the natural projection of M on N . Then we can take

h(r) = r + s, r ∈ R1 , s ∈ R2 , where s is either 0 or r. By replacing M by an arbitrary

direct sum of copies of R, we can produce two choices for the component of h(r) in each

Ri , i = 2, 3, . . . (with the restriction that only ¬nitely many components are nonzero).

Thus there will be in¬nitely many possible choices for h altogether.

31

Section 10.6

1. f g(a) = (g(a), 0) + W and g f (a) = (0, f (a)) + W , with (g(a), 0) ’ (0, f (a)) =

(g(a), ’f (a)) ∈ W . Thus f g = g f .

2. If (b, c) ∈ W , then b = g(a), c = ’f (a) for some a ∈ A. Therefore

g (c) + f (b) = ’g f (a) + f g(a) = 0

and h is well-de¬ned.

3. hg (c) = h((0, c) + W ) = g (c) + 0 = g (c) and hf (b) = h((b, 0) + W ) = 0 + f (b) =

f (b).

4. h ((b, c) + W ) = h ((0, c) + W + (b, 0) + W ) = h (g (c) + f (b)) = h g (c) + h f (b) =

g (c) + f (b) = h((b, c) + W ).

5. If f (b) = 0, then by de¬nition of f , (b, 0) ∈ W , so for some a ∈ A we have b = g(a)

and f (a) = 0. Since f is injective, a = 0, hence b = 0 and f is injective.

6. If b ∈ B, c ∈ C, then surjectivity of f gives c = f (a) for some a ∈ A. Thus

f (b + g(a)) = (b + g(a), 0) + W = (b + g(a), 0) + W + (’g(a), f (a)) + W [note

that (’g(a), f (a)) ∈ W ] = (b, f (a)) + W = (b, c) + W , proving that f is surjective.

7. If (a, c) ∈ D, then f (a) = g(c) and f g (a, c) = f (a), gf (a, c) = g(c). Thus f g = gf .

8. If x ∈ E, then f g (x) = gf (x), so (g (x), f (x)) ∈ D. Take h(x) = (g (x), f (x)),

which is the only possible choice that satis¬es g h = g and f h = f .

9. If (a, c) ∈ D and f (a, c) = 0, then c = 0, so f (a) = g(c) = 0. Since f is injective,

a = 0. Consequently, (a, c) = 0 and f is injective.

10. If c ∈ C, then there exists a ∈ A such that f (a) = g(c). Thus (a, c) ∈ D and

f (a, c) = c, proving that f is surjective.

11. If x, y ∈ I, then f (xy) = xf (y) and f (xy) = f (yx) = yf (x). Thus xf (y) = yf (x),

and if x and y are nonzero, the result follows upon division by xy.

12. We must extend f : I ’ Q to h : R ’ Q. Let z be the common value f (x)/x, x ∈ I,

x = 0. De¬ne h(r) = rz, r ∈ R. Then h is an R-homomorphism, and if x ∈ I, x = 0,

then h(x) = xz = xf (x)/x = f (x). Since h(0) = f (0) = 0, h is an extension of f and

the result follows from (10.6.4).

Section 10.7

1. The only nonroutine veri¬cation is the check that sf ∈ HomR (M, N ):

(sf )(rm) = f (rms) = rf (ms) = r[(sf )(m)].

2. (f r)(ms) = f (r(ms)) = f ((rm)s) = f (rm)s = [(f r)(m)]s.

3. (sf )(mr) = s(f (mr)) = s(f (m)r) = (sf (m)r = [(sf )(m)]r.

4. (f r)(sm) = (f (sm))r = (sf (m))r = s(f (m)r) = s[(f r)(m)].

32

5. In Problem 2, M and N are right S-modules, so we write f on the left: (f r)m =

f (rm). In Problem 3, M and N are right R-modules, and we write f on the left:

(sf )m = s(f m). In Problem 4, M and N are left S-modules, and we write f on the

right: m(f r) = (mf )r.

6. Let y ∈ M , r ∈ R with r = 0. By hypothesis, x = 1 y ∈ M , so we have x ∈ M such

r

that y = rx, proving M divisible.

7. If y ∈ M , r ∈ R with r = 0, we must de¬ne 1 y. Since M is divisible, there exists x ∈ M

r

such that y = rx, and we take 1 y = x. If x ∈ M and y = rx , then r(x ’ x ) = 0,

r

and since M is torsion-free, x = x . Thus x is unique and scalar multiplication is

well-de¬ned.

8. Let f be a nonzero R-homomorphism from Q to R. Then f (u) = 1 for some u ∈ Q. [If

f (x) = r = 0, then rf (x/r) = f (rx/r) = f (x) = r, so we can take u = x/r.] Now if s

is a nonzero element of R, then sf (u/s) = f (su/s) = f (u) = 1, so f (u/s) is an inverse

of s. Consequently, R is a ¬eld, contradicting the hypothesis.

Section 10.8

1. Let A = R[X] where R is any commutative ring. As an R-algebra, A is generated

by X, but A is not ¬nitely generated as an R-module since it contains polynomials of

arbitrarily high degree.

2. The bilinear map determined by (X i , Y j ) ’ X i Y j induces an R-homomorphism of

R[X] —R R[Y ] onto R[X, Y ], with inverse determined by X i Y j ’ X i — Y j .

3. Abbreviate X1 , . . . , Xn by X and Y1 , . . . , Ym by Y . Let A be a homomorphic image

of R[X] under f , and B a homomorphic image of R[Y ] under g. Then A —R B is a

homomorphic image of R[X] — R[Y ](∼ R[X, Y ] by Problem 2) under f — g.

=

4. If f : A ’ B is an injective R-module homomorphism, then by hypothesis, (1—f ) : S—R

A ’ S —R B is injective. Also by hypothesis,

(1 — (1 — f )) : M —S S —R A ’ M —S S —R B

is injective. Since M —S S ∼ M , the result follows.

=

5. Let f : A ’ B be injective. Since A —S S ∼ A and B —S S ∼ B, it follows from the

= =

hypothesis that (f — 1) : A —S (S —R M ) ’ B —S (S —R M ) is injective. Thus S —R M

is a ¬‚at S-module.

6. ± is derived from the bilinear map S ’1 R — M ’ S ’1 M given by (r/s, x) ’ rx/s. We

must also show that β is well-de¬ned. If x/s = y/t, then there exists u ∈ S such that

utx = usy. Thus

1 ut 1 1 1

—x= —x= — utx = — usy = — y

s sut sut sut t

as required. By construction, ± and β are inverses of each other and yield the desired

isomorphism.

7. We must show that S ’1 R —R is an exact functor. But in view of Problem 6, an

equivalent statement is the localization functor S ’1 is exact, and this has already been

proved in Section 8.5, Problem 5.

33

Section 10.9

1. The proof of (10.9.4) uses the fact that we are working in the category of modules. To

simply say “duality” and reverse all the arrows, we would need an argument that did

not depend on the particular category.

2. Let N be the direct limit of the Ni . The direct system {Ni , h(i, j)} induces a direct

system {M — Ni , 1 — h(i, j)}. Compatibility in the new system reduces to compatibility

in the old system; tensoring with 1 is harmless. Since compatible maps fi : Ni ’ B

can be lifted to f : N ’ B, it follows that compatible maps gi : M — Ni ’ B can be

lifted to g : M — N ’ B. Thus M — N satis¬es the universal mapping property for

{M — Ni }.

3. The direct limit is A = ∪∞ An , with ±n : An ’ A the inclusion map

n=1

5. Each R-homomorphism f from the direct sum of the Ai to B induces an

R-homomorphism fi : Ai ’ B. [fi is the injection of Ai into the direct sum, fol-

lowed by f ]. Take ±(f ) = (fi , i ∈ I) ∈ i HomR (Ai , B). Conversely, given such a

family (fi , i ∈ I), the fi can be lifted uniquely to β(fi , i ∈ I) = f . Since ± and β are

inverse R-homomorphisms, the result follows.

6. If f : A ’ i Bi , de¬ne ±(f ) = (pi f, i ∈ I) ∈ i HomR (A, Bi ), where pi is the

projection of the direct product onto the ith factor. Conversely, given (gi , i ∈ I), where

gi : A ’ Bi , the gi can be lifted to a unique g : A ’ i Bi such that pi g = gi for all i.

If we take β(gi , i ∈ I) = g, then ± and β are inverse R-homomorphisms, and the result

follows.

7. There is a free module F such that F = M • M , and since F is torsion-free, so is M .

Since M is injective, it is divisible, so by Problem 7 of Section 10.7, M is a vector

space over the quotient ¬eld Q, hence a direct sum of copies of Q. Therefore, using

Problem 5 above and Problem 8 of Section 10.7,

HomR (M, R) = HomR (•Q, R) ∼ Hom(Q, R) = 0.

=

8. By Problem 7, HomR (M, R) = 0. Let M be a direct summand of the free module F

with basis {ri , i ∈ I}. If x is a nonzero element of M , then x has some nonzero

coordinate with respect to the basis, say coordinate j. If pj is the projection of F on

coordinate j (the j th copy of R), then pj restricted to M is a nonzero R-homomorphism

from M to R. (Note that x does not belong to the kernel of pj .) Thus the assumption

that M = 0 leads to a contradiction.

BIBLIOGRAPHY

General

Cohn, P.M., Algebra, Volumes 1 and 2, John Wiley and Sons, New York, 1989

Dummit, D.S. and Foote, R.M., Abstract Algebra, Prentice-Hall, Upper Saddle River,

NJ, 1999

Hungerford, T.M., Algebra, Springer-Verlag, New York, 1974

Isaacs, I.M., Algebra, a Graduate Course, Brooks-Cole, a division of Wadsworth, Inc.,

Paci¬c Grove, CA, 1994

Jacobson, N., Basic Algebra I and II, W.H. Freeman and Company, San Francisco, 1980

Lang, S., Algebra, Addison-Wesley, Reading, MA, 1993

Modules

Adkins, W.A., and Weintraub, S.H., Algebra, An Approach via Module Theory,

Springer-Verlag, New York, 1992

Blyth, T.S., Module Theory, Oxford University Press, Oxford, 1990

Basic Group Theory

Alperin, J.L., and Bell, R.B., Groups and Representations, Springer-Verlag, New York,

1995

Humphreys, J.F., A Course in Group Theory, Oxford University Press, Oxford 1996

Robinson, D.S., A Course in the Theory of Groups, Springer-Verlag, New York, 1993

Rose, J.S., A Course on Group Theory, Dover, New York, 1994

Rotman, J.J., An Introduction to the Theory of Groups, Springer-Verlag, New York, 1998

Fields and Galois Theory

Adamson, I.T., Introduction to Field Theory, Cambridge University Press, Cambridge,

1982

Garling, D.J.H., A Course in Galois Theory, Cambridge University Press, Cambridge,

1986

Morandi, P., Fields and Galois Theory, Springer-Verlag, New York, 1996

Roman, S., Field Theory, Springer-Verlag, New York, 1995

Rotman, J.J., Galois Theory, Springer-Verlag, New York, 1998

Algebraic Number Theory

Borevich, Z.I., and Shafarevich, I.R., Number Theory, Academic Press, San Diego, 1966

Fr¨hlich, A., and Taylor, M.J., Algebraic Number Theory, Cambridge University Press,

o

Cambridge, 1991

Gouvea, F.Q., p-adic Numbers, Springer-Verlag, New York, 1997

Janusz, G.J., Algebraic Number Fields, American Mathematical Society, Providence,

1996

Lang, S., Algebraic Number Theory, Springer-Verlag, New York, 1994

Marcus, D.A., Number Fields, Springer-Verlag, New York, 1977

Samuel, P., Algebraic Theory of Numbers, Hermann, Paris, 1970

Algebraic Geometry and Commutative Algebra

Atiyah, M.F., and Macdonald, I.G., Introduction to Commutative Algebra,

Addison-Wesley, Reading, MA, 1969

Bump, D., Algebraic Geometry, World Scienti¬c, Singapore, 1998

Cox, D., Little, J., and O™Shea, D., Ideals, Varieties, and Algorithms, Springer-Verlag,

New York, 1992

Eisenbud, D., Commutative Algebra with a View Toward Algebraic Geometry,

Springer-Verlag, New York, 1995

Fulton, W., Algebraic Curves, W.A. Benjamin, New York, 1969

Hartshorne, R., Algebraic Geometry, Springer-Verlag, New York, 1977

Kunz, E., Introduction to Commutative Algebra and Algebraic Geometry, Birkh¨user,a

Boston, 1985

Matsumura, H., Commutative Ring Theory, Cambridge University Press, Cambridge,

1986

Reid, M., Undergraduate Algebraic Geometry, Cambridge University Press, Cambridge,

1988

Reid, M., Undergraduate Commutative Algebra, Cambridge University Press,

Cambridge, 1995

Shafarevich, I.R., Basic Algebraic Geometry, Volumes 1 and 2, Springer-Verlag, New

York 1988

Ueno, K., Algebraic Geometry 1, American Mathematical Society, Providence 1999

Noncommutative Rings

Anderson, F.W., and Fuller, K.R., Rings and Categories of Modules, Springer-Verlag,

New York, 1992

Beachy, J.A., Introductory Lectures on Rings and Modules, Cambridge University Press,

Cambridge, 1999

Farb, B., and Dennis, R.K., Noncommutative Algebra, Springer-Verlag, New York, 1993

Herstein, I.N., Noncommutative Rings, Mathematical Association of America,

Washington, D.C., 1968

Lam, T.Y., A First Course in Noncommutative Rings, Springer-Verlag, New York, 1991

Group Representation Theory

Curtis, C.W., and Reiner, I., Methods of Representation Theory, John Wiley and Sons,

New York, 1981

Curtis, C.M., and Reiner, I., Representation Theory of Finite Groups and Associative

Algebras, John Wiley and Sons, New York, 1966

Dornho¬, L., Group Representation Theory, Marcel Dekker, New York, 1971

James, G., and Liebeck, M., Representations and Characters of Groups, Cambridge

University Press, Cambridge, 1993

Homological Algebra

Hilton, P.J., and Stammbach, U., A Course in Homological Algebra, Springer-Verlag,

New York, 1970

Hilton, P., and Wu, Y-C., A Course in Modern Algebra, John Wiley and Sons, New

York, 1974

Mac Lane, S., Categories for the Working Mathematician, Springer-Verlag, New York,

1971

Rotman, J.J., An Introduction to Algebraic Topology, Springer-Verlag, New York, 1988

Rotman, J.J., An Introduction to Homological Algebra, Springer-Verlag, New York, 1979

Weibel, C.A., An Introduction to Homological Algebra, Cambridge University Press,

Cambridge, 1994

List of Symbols

Throughout the text, ⊆ means subset, ‚ means proper subset

Zn integers modulo n 1.1

Z integers 1.1

A subgroup generated by A 1.1

Sn symmetric group 1.2

An alternating group 1.2

D2n dihedral group 1.2

• Euler phi function 1.1, 1.3

normal subgroup 1.3

proper normal subgrouad 1.3

ker kernel 1.3, 2.2

∼ isomorphism 1.4

=

Z(G) center of a group 1.4

H —K direct product 1.5

Q rationals 2.1

Mn (R) matrix ring 2.1

R[X] polynomial ring 2.1

R[[X]] formal power series ring 2.

End endomorphism ring 2.1

X ideal generated by X 2.2

UFD unique factorization domain 2.6

PID principal ideal domain 2.6

ED Euclidean domain 2.7

min(±, F ) minimal polynomial 3.1

i Ki composite of ¬elds 3.1

Gal(E/F ) Galois group 3.5

the module {0} and the ideal {0}

0 4.1

•i Mi direct sum of modules 4.3

i Mi sum of modules 4.3

HomR (M, N ) set of R-module homomorphisms from M to N 4.4

EndR (M ) endomorphism ring 4.4

g•x group action 5.1

G commutator subgroup 5.7

G(i) derived subgroups 5.7

S|K presentation of a group 5.8

F(H) ¬xed ¬eld 6.1

G(K) ¬xing group 6.1

GF (pn ) ¬nite ¬eld with pn elements 6.4

nth cyclotomic polynomial

Ψn (X) 6.5

∆ product of di¬erences of roots 6.6

D discriminant 6.6, 7.4

N [E/F ] norm 7.3

T [E/F ] trace 7.3

char characteristic polynomial 7.3

nP (I) exponent of P in the factorization of I 7.7

vp p-adic valuation 7.9

| |p p-adic absolute value 7.9

V (S) variety in a¬ne space 8.1

I(X) ideal of a set of points 8.1

√ 1 , . . . , Xn ]

k[X polynomial ring in n variables over the ¬eld k 8.1

I radical of an ideal 8.3

k(X1 , . . . , Xn ) rational function ¬eld over k 8.4

S ’1 R localization of the ring R by S 8.5

S ’1 M localization of the module M by S 8.5

N (R) nilradical of the ring R 8.6

M —R N tensor product of modules 8.7

UMP universal mapping property 8.7

A—B tensor (Kronecker) product of matrices 8.7

kG group algebra 9.5

RG group ring 9.5

J(M ), J(R) Jacobson radical 9.7

HomR (M, ), HomR ( , N ) hom functors 10.3

M —R , —R N tensor fuctors 10.3

Q/Z additive group of rationals mod 1 10.6

(also 1.1, Problem 7)

Z(p∞ ) quasicyclic group A10

G[n] elements of G annihilated by n A10

Hn homology functor S1

fg chain homotopy S1

‚ connecting homomorphism S2, S3

P— ’ M projective resolution S4

M ’ E— injective resolution S4

Ln F left derived functor S5

Rn F right derived functor S5

derived functor of —

Tor S5

Ext derived functor of Hom S5

INDEX

abelian category 10.4 character 6.1

characteristic of a ring or ¬eld 2.1

abelian group 1.1

characteristic polynomial 7.3

absolute value 7.9

characteristic subgroup 5.7

action of a group on a set 5.1

chief series 5.6

adjoint associativity 10.7

Chinese remainder theorem 2.3

adjoint functors 10.7

class equation 5.2

a¬ne n-space 8.1

cokernel 10.1

a¬ne variety 8.1

colorings 5.3

AKLB setup 7.3

commutative diagram 1.4

algebra 4.1

commutative ring 2.1

algebraic closure 3.3, 10.9

commutator 5.7

algebraic curve 8.3

compatible morphisms 10.9

algebraic element 3.1

complete ring of fractions 2.8

algebraic extension 3.1

composite of ¬elds 3.1, 6.2

algebraic function ¬eld 6.9

composition factors 5.6, 7.5

algebraic geometry 8.1¬

composition length 5.6, 7.5

algebraic integers 7.1

composition of morphisms 10.1

algebraic number 3.3, 7.1

composition series 5.6, 7.5

algebraic number theory 7.1¬, 7.3

conjugate elements 5.1, 5.2

algebraically closed ¬eld 3.3

conjugate sub¬elds 6.2

algebraically independent set 6.9

conjugate subgroups 5.1, 6.2

algebraically spanning set 6.9

conjugates of a ¬eld ement 3.5

alternating group 1.2

conjugation 5.1, 5.2-1

annihilator 4.2, 9.2, 9.7

connecting homomorphism S2, S3

archimedian absolute value 7.9

constructible numbers and points 6.8

Artin-Schreier theorem 6.7

content 2.9

Artinian modules 7.5

contravariant functor 10.3

Artinian rings 7.5

coproduct 10.2

ascending chain condition (acc) 2.6, 7.5

core 5.1

associates 2.6

correspondence theorem for

associative law 1.1, 2.1

groups 1.4

automorphism 1.3

correspondence theorem for

Baer™s criterion 10.6 modules 4.2

base change 10.8 correspondence theorem for

basis 4.3 rings 2.3

bilinear mapping 8.7 coset 1.3

binomial expansion modulo p 3.4 counting two ways 5.3

binomial theorem 2.1 covariant functor 10.3

boundary S1 cycle 1.2, S1

cyclic extension 6.7

canonical map 1.3

cyclic group 1.1

category 10.1 cyclic module 4.2, 9.1, 9.2, 9.7

Cauchy™s theorem 5.4 cyclotomic extension.5

Cayley™s theorem 5.1 cyclotomic ¬eld 6.5, 7.2

center of a group 1.4 cyclotomic polynomial 6.5

center of a ring 4.1

central series 5.7 decomposable module 9.6

centralizer 5.2 Dedekind domain 7.6, 7.7

chain complex S1 Dedekind™s lemma 6.1, 6.7, 7.3, 7.4

chain homotopy S1 degree 2.5

chain map S1 deleted projective (or injective)

chain rule 1.3, 3.1 resolution S4

derivative of a polynomial 3.4 faithful action 5.1

derived functors S5 faithful module 7.1, 9.2, 9.4

derived length 5.7 faithful representation 9.5

derived series 5.7 Fermat primes 6.8

descending chain condition 7.5 Fermat™s little theorem 1.3

diagram chasing 4.7 ¬eld 2.1

di¬erential S1 ¬eld discriminant 7.4

dihedral group 1.2, 5.8, 5.8 ¬nite abelian groups 4.6

(in¬nite dilhedral group), 9.5 ¬nite extension 3.1

direct limit 10.9 ¬nite ¬elds 6.4

direct product of groups 1.5 ¬nitely cogenerated module 7.5

direct product of modules 4.3 ¬nitely generated algebra 10.8

direct product of rings 2.3 ¬nitely generated module 4.4

direct sum of modules 4.3 ¬nitely generated submodule 7.5

direct system 10.9 ¬ve lemma 4.7

directed set 10.9 ¬xed ¬eld 6.1

discriminant 6.6, A6, 7.4 ¬xing group 6.1

divides means contains 2.6, 7.7 ¬‚at modules 10.8

divisible abelian group A10 forgetful functor 10.3

divisible module 10.6 formal power series 2.1, 8.2

division ring 2.1, 9.1 four group 1.2, 1.5, A6

double centralizer 9.2 four lemma 4.7

double dual functor 10.3 fractional ideal 7.6

dual basis 7.4 Frattini argument 5.8-2

duality 10.1 free abelian gup functor 10.3

duplicating the cube 6.8 free group 5.8-1

free module 4.3,15

Eisenstein™s irreducibility criterion 2.9

free product 10.2-2

elementary divisors 4.6

Frobenius automorphism 3.4, 6.4

elementary symmetric functions 6.1

full functor 10.3

embedding 3.3, 3.5

full ring of fractions 2.8

embedding in an injective module 10.7

full subcategory 10.3

endomorphism 1.3, 4.4

functor 10.3

epic 10.1

fundamental decomposition theorem

epimorphism 1.3

(for ¬nitely generated modules

equivalent absolute values 7.9

over a PID) 4.6

equivalent matrices 4.4

fundamental theorem of Galois theory

equivalent matrix

6.2-1

representations 9.5

Euclidean domain 2.7

Galois extension 3.5, 6.1¬.

Euler™s identity 2.1

Galois group 3.5, 6.1¬

Euler™s theorem 1.3

Galois group of a cubic, 6.6

evaluation map 2.1

Galois group of a polynomial 6.3

exact functor 8.5, 10.4

Galois group of a quadratic 6.3

exact sequence 4.7

Galois group of a quartic A6

exponent of a group 1.1, 6.4

Gauss™ lemma 2.9

Ext S5

Gaussian integers 2.1, 2.7

extension of a ¬eld 3.1

general equation of degree n 6.8

extension of scalars 8.7, 10.8

general linear group 1.3

exterior algebra 8.8

generating set 4.3

generators and relations 1.2, 4.6, 5.8

F-isomorphism, etc. 3.2

greatest common divisor 2.6, 7.7

factor theorem for groups 1.4

group 1.1

factor theorem for modules 4.2

factor theorem for rings 2.3 group algebra 9.5

group representations 9.5 Jacobson radical 9.7

group ring 9.5 Jacobson™s theorem 9.2

Jordan-Holder theorem 5.6, 7.5

Hermite normal form 4.5

Hilbert basis theorem 8.2 kernel 1.3, 2.2, 10.1

Hilbert™s Nullstellensatz 8.3, 8.4 kernel of an action 5.1

Hilbert™s Theorem 90 7.3 Kronecker product of matrices 8.7

hom functors 10.3-1 Krull-Schmidt theorem 9.6

homology functors S1 Kummer extension 6.7

homology group S1

Lagrange interpolation formula 2.5

homology module S1

Lagrange™s theorem 1.3

homomorphism from R to M deter-

Laurent series 7.9

mined by what it does to the

leading coe¬cient 2.5

identity, 9.4, S6

least common multiple 2.6, 7.7

homomorphism of algebras 4.1

left adjoint 10.7

homomorphism of groups 1.3

left cancellable 10.1

homomorphism of modules 4.1

left derived functors S5

homomorphism of rings 2.2

left exact functor 10.4

Hopkins-Levitzki theorem 9.8

left ideal 2.2

hypersurface 8.2

left resolution S4

ideal 2.2, 8.1 left-Noetherian ring 9.8

ideal class group 7.8 left-quasiregular element 9.7

idempotent linear transformation 9.5 left-semisimple ring 9.6

image 2.3, 4.1 length of a module 7.5

indecomposable module 9.6 lifting of a map 4.3, 10.2

index 1.3 linearly indepdent set 4.3

inductive limit 10.9 local ring 2.4, 7.9, 8.5

initial object 10.1 localization 2.8, 8.5

injection (inclusion) 4.7

long division 6.4

injective hull 10.7

long exact homology sequence S3

injective modules 10.6

injective resolution S4 Maschke™s theorem 9.6

inner automorphism 1.4, 5.7 matrices 2.1, 4.4

integral basis 7.2, 7.4 maximal ideal 2.4, 8.3

integral closure 7.1 maximal submodule 9.7

integral domain 2.1 metric 7.9

integral extensions 7.1 minimal generating set 9.8

integral ideal 7.6 minimal left ideal 9.3

integrally closed 7.1 minimal polynomial 3.1

invariant factors 4.5 minimal prime ideal 8.4

inverse limit 10.9 modding out 5.7

inverse system 10.9 modular law 4.1

inversions 1.2 module 4.1

irreducible element 2.6 modules over a principal ideal domain

irreducible ideal 8.6 4.6, 10.5

irreducible polynomial 2.9 monic 10.1

irreducible variety 8.1 monoid 1.1

isomorphic groups 1.1 monomorphism 1.3

isomorphism 1.3 morphism 10.1

isomorphism extension theorem 3.2

Nakayama™s lemma 9.8

isomorphism theorems for groups 1.4

natural action 5.3, 6.3

isomorphism theorems for modules 4.2

isomorphism theorems for rings 2.3 natural map 1.3

natural projection 4.7, 7.5, 9.2, 9.4, 9.5, principal ideal domain 2.6

10.5 product 10.2

natural transformation 10.3 product of an ideal and a module 9.3

naturality S3 product of ideals 2.3, 8.5, 7.6

Newton™s identities A6 projection 4.7, 9.2, 9.4, 9.5, 10.5

nil ideal 9.7 projection functor 10.3

nilpotent element 8.6, 9.7 projective basis lemma 10.5

nilpotent group 5.7 projective limit 10.9

nilpotent ideal 9.7 projective modules 9.8, 10.5

nilradical 8.6 projective resolution S4

Noetherian modules 4.6, 7.5 proper ideal 2.2

Noetherian rings 4.6, 7.5 Prufer group A10

nonarchimedian absolute value 7.9 pullback 10.6

noncommuting indeterminates 9.8 purely inseparable 3.4

nontrivial ideal 2.2 pushout 10.6,3

norm, 7.1, 7.3

quadratic extensions 6.3, 7.2

normal closure, 3.5

quasicyclic group A10

normal extension 3.5

quasi-regular element 9.7

normal series 5.6

quaternion group 2.1

normal Sylow p-subgroup 5.5

quaternions 2.1

normalizer 5.2

quotient ¬eld 2.8

Nullstellensatz 8.3, 8.4

quotient group 1.3

number ¬eld 7.1

quotient ring 2.2

objects 10.1

R-homomorphism on R 9.4

opposite category 10.1

Rabinowitsch trick 8.4

opposite ring 4.4

radical extension 6.8

orbit 5.2

radical of an ideal 8.3

orbit-counting theorem 5.3

rank of a free module 4.4

orbit-stabilizer theorem 5.2

rational integer 7.2

order 1.1

rational root test 2.9

order ideal 4.2

rationals mod 1 1.1 (Problem 7), 10.4,

orthogonal idempotents 9.6

10.6, A10

Ostrowski™s theorem 7.9

re¬nement 5.6

p-adic absolute value 7.9 regular action 5.1, 5.2

p-adic integers 7.9 regular n-gon 6.8

p-adic numbers 7.9 regular representation 9.5

p-adic valuation 7.9 relatively prime ideals 2.3

p-group 5.4 remainder theorem 2.5

perfect ¬eld 3.4 representation 9.5

permutation 1.2 residue ¬eld 9.8

permutation group 1.2 resolvent cubic A6

permutation module 9.5 restriction of scalars 8.7, 10.8

polynomial rings 2.1, 2.5 right adjoint 10.7

polynomials over a ¬eld 3.1 right cancellable 10.1

power sums, A6 right derived functors S5

preordered set 10.2 right exact functor 10.4

primary component A10 right ideal 2.2

primary decomposition 8.6 right resolution S4

primary ideal 8.6 right-Noetherian ring 9.8

prime element 2.6hbn right-quasiregular element 9.7

prime ideal 2.4 right-semisimple ring 9.6

primitive element, theorem of 3.5, 6.6 ring 2.1

primitive polynomial 2.9 ring of fractions 2.8, 8.5

torsion subgroup A10

Schreier re¬nement theorem 5.6

torsion submodule 4.6

Schur™s lemma 9.2

torsion-free module 4.6

semidirect product 5.8

trace 7.3

semigroup 1.1

transcendence basis 6.9

semisimple module 9.1

transcendence degree 6.9

semisimple ring 9.3

transcendental element 3.1

separable element 3.4

transcendental extension 6.9

separable extension 3.4

transcendental number 3.3

separable polynomial 3.4

transitive action 5.2

series for a module 7.5

transitive subgroup of Sn 6.3

simple group 5.1, 5.5

transitivity of algebraic extensions 3.3

simple left ideal 9.3

transitivity of separable extensions 3.4

simple module 7.5, 9.1, 9.2

transitivity of trace and norm 7.3

simple ring 9.3, 9.5

transposition 1.2

simplicity of the alternating group 5.6

trisecting the angle 6.8

simultaneous basis theorem 4.6

trivial absolute value 7.9

skew ¬eld 2.1

trivial action 5.1, 5.2

Smith normal form 4.5

twisted cubic 8.3

snake diagram S2

two-sided ideal 2.2

snake lemma S2

solvability by radicals 6.8 ultrametric inequality 7.9

solvable group 5.7 underlying functor 10.3

spanning set 4.3, 6.9 unimodular matrix 7.4

special linear group 1.3 unique factorization domain 2.6

split exact sequence 4.7, 5.8 unique factorization of ideals 7.7

splitting ¬eld 3.2 unit 2.1

squaring the circle 6.8 universal mapping property (UMP)

standard representation of a p-adic 8.7, 10.2

integer 7.9 upper central series 5.7

Steinitz exchange 6.9

valuation 7.9

Stickelberger™s theorem 7.4

valuation ideal 7.9

subcategory 10.3

valuation ring 7.9

subgroup 1.1

Vandermonde determinant A6, 7.4

submodule 4.1

vector space as an F [X]-module 4.1

subnormal series 5.6

Von Dyck™s theorem 5.8

subring 2.1

sum of ideals 2.2

weak Nullstellensatz 8.3

sum of modules 4.3

Wedderburn structure theorem 9.5

Sylow p-subgroup 5.4

Wedderburn-Artin theorem 9.4

Sylow theorems 5.4

symmetric group 1.2, 6.6 Zariski topology 8.1

symmetric polynomial 6.1 Zassenhaus lemma 5.6

zero object 10.1

tensor functors 10.3

tensor product of matrices 8.7

tensor product of module homo-

morphisms 8.7

tensor product of modules 8.7

terminal object 10.1

Tor S5

torsion abelian group 8.7 (Problem 3),

10.2

torsion element 4.6

torsion module 4.6