1. If
RM is simple, then
M =
Rx for any 0 ≠
x M. Hence
α :
R →
M given by
α(
r) =
rx is an onto
R-linear map, so
M ≅
R/
L where
L =
ker(
α), and
L is maximal because
M is simple. Conversely,
R/
L is simple for every maximal left ideal
L. Conversely,
R/
L is simple for every maximal left ideal
L again by Theorem 6 §8.1.
3. Define σ : Ra → Rb by σ(ra) = rb. This is well defined because ra = 0 implies rbR = r(aR) = 0, so rb = 0 . So σ is an onto homomorphism of left R-modules with σ(a) = b. Finally if σ(ra) = 0 then rb = 0 so (ra)R = r(aR) = r(bR) = 0, so ra = 0 . Hence σ is one-to-one.
5. If
L1⊆
L2 ⊆
are left ideals of
eRe, then
RL1⊆
RL2 ⊆
are left ideals of
R so
RLn =
RLn+1 =
for some
n by hypothesis. If
i ≥
n, the fact that
Li ⊆
eRe for each
i, gives
Li =
eLi ⊆
eRLi =
eRLn =
eReLn ⊆
Ln. Hence
Ln =
Ln+1 =
, as required.
7. If RM is finitely generated, then 143 M is an image of Rn for some n ≥ 1 by Theorem 5 §7.1 and its Corollaries. Since Rn is noetherian as a left R-module (Corollary 1 of Lemma 2), the same is true of M by Lemma 2.
8. a. Observe that
ann(
X) is a left ideal for any
X ⊆
M (note that ann (∅) =
R) . Hence use the artinian condition to choose
ann(
X) minimal in
Then
ann(
M) ⊆
ann(
X) is clear; we prove equality. If
a ann(
X), suppose that
a ∉
ann(
M), say
am ≠ 0 with
m M. Put
Y =
X ∪ {
m} . Then
ann(
Y) ⊆
ann(
X), so
ann(
Y) =
ann(
X) by minimality. But
a ann(
Y) and so
am = 0, a contradiction. Hence
a ann(
M) and we have proved (a).
9. Let
and
p does not divide
m} . If
where
Y is a subgroup of
X, choose
with
n maximal (
n exists because
Y ≠
X,
and
Then
by Theorem 4 §1.2 because
m and
pn are relatively prime. Hence
we claim that this is equality. To see this, let
where
p does not divide
m′ . Then
k ≤
n by the maximality of
n, so
whence
11. a. If
x M then
x−
π(
x)
ker
π, and it follows that
M =
π(
M)+ ker
π. If
y π(
M) ∩ ker
π write
y =
π(
z),
z M. Since
y ker
π, we have 0 =
π(
y) =
π2(
z) =
π(
z) =
y, so
π(
M)∩ ker
π = 0 . This shows that
M =
π(
M)⊕ ker
π. (c). We have ker
π = (1 −
π)(
M) because
π(
x) = 0 if and only if (1 −
π)(
x) =
x. And
π(
M) =
ker(1 −
π) because (1 −
π)(
x) = 0 if and only if
x =
π(
x).
13. a. Suppose that
K =
K1⊕
K2 ⊕
is an infinite direct sum of nonzero submodules of
M. Then
K⊃
K2 ⊕
K3 ⊕
⊃
K3 ⊕
K4 ⊃
contradicts the DCC, and
K1⊂
K1 ⊕
K2 ⊂
K1 ⊕
K2 ⊕
K3 ⊂
contradicts the ACC.
15. Here
and we consider
Then
e2 =
e and
is not simple as
is a submodule. But end
as rings.
17. a. We have
ker(
α)⊆
ker(
α2) ⊆
ker(
α3) ⊆
so, since
M is noetherian, there exists
n ≥ 1 such that
ker(
αn) =
ker(
αn+1) =
. If
x ker(
α) then (since
αn is onto) write
x =
αn(
y) for some
y M. Then 0 =
α(
x) =
αn+1(
y), so
y ker(
αn+1) =
ker(
αn) . Hence
x =
αn(
y) = 0, proving that
ker(
α) = 0.
19.
a. The easy verification that
θ is an
F-linear homomorphism of additive groups is left to the reader. To show that
θ is one-to-one, suppose that
θ(
r) = 0 ; that is,
uir = 0 for each
i. If 1 = ∑
iaiui, then
r = 1 ·
r = ∑
iaiuir = 0, so ker
θ = 0 and
θ is one-to-one. Finally, let
θ(
s) = [
sij] so that
Then:
Thus
and the proof is complete.
c.
c.